United States
Environmental Protection
Agency
A Short-term Test Method for
Assessing the
Reproductive Toxieity of
Endocrine-Disrupting
Chemicals Using the
Fathead Minnow
(Pimephales promotes)
-------
-------
EPA/600/R-01/067
June 2002
A Short-term Test Method for Assessing the Reproductive Toxicity of
Endocrine-Disrupting Chemicals Using the Fathead Minnow (Pimephales promelas)
U.S. Environmental Protection Agency
Office of Research and Development
National Health and Environmental Effects Research Laboratory
Mid-Continent Ecology Division
Duluth, MN 55804
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free.
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the U.S. Environmental
Protection Agency. It has been subjected to review by the National Health and Environmental
Effects Research Laboratory and approved for publication. Approval does not signify that the
contents reflect the views of the Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
u
-------
FOREWORD
i , ' !.
1 s ,
,. ' - I
This manual describes a test method with the fathead minnow (Pimephales promelas) suitable
for assessing potential reproductive effects of chemicals, with an emphasis on endocrine
pathways controlled by estrogens and androgens. The test is conducted with reproductively-
mature animals for 21 d. Endpoints assessed include: adult survival, reproductive behavior,
secondary sex characteristics, gonadosomatic index, gonadal histology, plasma concentrations of
vitellogenin and sex steroids (p-estradiol, testosterone, 11-ketotestosterone), fecundity, fertility,
and, if desired, FH viability. In addition to describing the test method, guidance is presented as to
interpretation of test results with respect to identification of potential endocrine-disrupting
chemicals. ,
This document can be cited as EPA/600/R-01/067, U.S. Environmental Protection Agency,
Office of Research and Development, National Health and Environmental Effects Research
Laboratory, Mid-Continent Ecology Division, Duluth, MN, USA.
in
-------
CONTENTS
i Page
Disclaimer ii
Foreword ,... iii
Contents — iv
Figures ... ix
Tables „.. x
Acknowledgments xi
A. Scope 1
B. Introduction | 2
1. Basic Reproductive Biology 2
2. Toxicology 5
C. Acronyms and Definitions 7
D. Principle of the Test 10
E. Information on the Test Chemical ! 11
1. Toxicity to Fish 11
2. Physico-Chemical Properties and Chemical Delivery 15
3. Range-Finding Toxicity Tests 17
4. Analytical Determinations 18
F. Validity of the Test , 18
G. Description of the Method 19
iv
-------
1. Exposure Apparatus '». 19
2. Test Chambers ...../ 20
3. Selection of Test Species 22
4. Test Water 22
5. Testtype 23
6. Stock Solutions (Aqueous Exposures) j...... 24
a. Solid-Liquid (slow-stir) Saturator 24
b. Liquid-Liquid (slow-stir) Saturator 26
c. Glass Wool Column Saturator ........> 28
d. Solvent Carriers 30
!' i _ • . .
7. Number of Treatments » 30
8. Replication 30
9. Performance Standard Test 33
10. Reference Toxicant Tests 34
H. Procedure ....< i 35
1. Pre-Exposure Reproduction w 35
a. Selection of Fish '.. ; 35
b. Conditions 36
i
c., Observations 36
d. Suitability for Testing , 37
2. Conditions of the 21-d Reproduction Test. J.' 38
a^ Duration ,- • • •..-,. - • 38
-------
b. Loading ; 38
c. Light and Temperature ' 39
d. Feeding 39
e. Test Concentrations 40
f. Controls 42
g. Fish : 43
h. Cleaning of Test Chambers , 44
3. Test Chemical Administration 44
a. Flow-Through Aqueous-Delivery .... i 44
b. Flow-Through Aqueous/Solvent-Delivery 45
c. Intraperitoneal (i.p.) Injection 45
d. Dietary Exposure 47
4. Analytical Determinations and Measurements 49
a. Test Chemical 49
b. Water Characteristics ! 50
5. Measurement of Test Endpoints : 51
i . i
a. Survival of Adults 51
b. Behavior of Adults 51
c. Fecundity 52
d. Fertility 53
e. Embryo Hatch .; 53
f. Larval Survival and Morphology 54
vi
-------
g. Secondary Sex Characteristics :..' 55
h. Blood Collection • • 56
i. Vitellogenin : 57
j. Sex Steroids • 58
k. Gonadosomatic Index — .i 59
1. Gonadal Histology •....>. .-.'* 60
m. Tissue Residues ; 62
n. Other Endpoints ^ 63
I. Treatment of Results ... .....'.'. -66
1. Overview 66
2. Analysis of Fecundity Data , 67
a. Data Preparation and Adjustment for Mortality 67
b. Choice of Analysis .-.i : 69
c. Analysis of Variance 73
d. Jonckherre-Terpstra Test of Ordered Alternatives 74
e. Kruskall-Wallis Test of Equal Treatment Medians ... i 78
f. Nonparametric Multiple Comparisons.Between Treatments '. 78
3. Analysis of Other Endpoints 80
a. Data Preparation :. 80
b. Outliers 80
c. Choice of Analysis i •• 81
d. Summarization 85
vii'
-------
J. TestReport 85
i
1. Test Chemical 85
2. Test Animals 86
3. Test Conditions 86
4. Results 87
K. Quality Assurance 87
L. Interpretation of Results 90
M. References ! 98
N. Appendices 108
A. Data Reporting Forms 108
B. Measurement of Plasma Vitellogenin in Fathead Minnows by Competitive ROSA ..116
C. Measurement of Plasma Steroid Concentrations in Fathead Minnows by RIA 125
D. Histological Techniques for Fathead Minnow Gonads 135
E. Histological Evaluation of Fathead Minnow Gonads 139
-------
FIGURES
Figure
Page
1. Mature male (left) and female (right) fathead minnow, Pimephales promelas 3
2. Liquid-liquid (slow-stir) saturation unit
27
3. Glass wool column saturation unit : 29
4. Number of replicates required per treatment to detect differences between treatment
means with ANOVA using three treatments when a = 0.05 and 0 = 0.10 or 0.20.
Ratios are the desired minimum detectable difference in treatment means divided by
the pooled standard deviation
32
5. Handling technique for administering test chemicals to fathead minnows by
i.p. injection •...*'..
47
6. Fathead minnow blood collection process ; 57
7. Flow chart for statistical analysis of data from the 21-d reproduction test with
the fathead minnow
84
-------
TABLES
Table
1. Overview of recommended exposure conditions for the fathead minnow 21-d
reproduction test ! 21
2. Recommended ranges of water quality characteristics for fathead minnow reproduction . 23
3. P-values corresponding to specified values Of /, the test statistic for the Jonckherre-
Terpstra test of ordered alternatives (Daniel 1978). Values are provided for n=4
replicates of k=3 treatments, n=3 replicates of k=4 treatments, and n=4 replicates of k=4
treatments. Values of J for n=4, k=4 are included only when the p-value brackets a
commonly used a level (e.g., the two values J=66 and J=67 bracket the 0.05 p-value). .. 77
4. Critical values for the Kruskal-Wallis test (Daniel 1978) and Scheffe's nonparametric
multiple comparisons approach with experiment-wise error rates of 5% and 1%.
Values are provided for n=4 replicates of fc=3 treatments, n=3 replicates of 7c=4
• treatments, and n=4 replicates of k=4 treatments
79
5. Performance of fathead minnow screening tests with known endocrine-disrupting
chemicals 96
-------
ACKNOWLEDGMENTS
The basic test method described in this document is derived from research performed by
scientists at the Mid-Continent Ecology Division (MED) of the National Health and
Environmental Effects Research Laboratory (NHEERL) in the Office of Research and
Development (ORD) of the U.S. Environmental Protection Agency (U.S. EPA). Portions of the
document were prepared via a contract (8D-1337-NAEX) with the University of Wisconsin-
Superior (Dr. Daniel Call, Mr. Larry Brooke). Histological guidance and related appendices
I1 i • •
were provided by Dr. Richard Leino, University of Minnesota-Duluth School of Medicine (JJ-
0139-NTTX). Dr. Philip Dixon, Iowa State University, contributed to the statistical guidance
(OJ-0248-NAEX). The lead authors of the document from MED are Dr. Gerald Ankley and Ms.
Kathleen Jensen, with significant input by Dr. Michael Hornung, Mr. Michael Kahl, Mr. Joseph
Korte, and Ms. Elizabeth Makynen. The authors gratefully acknowledge review comments from
Dr. Tala Henry, Experimental Toxicology Division, NHEERL; and Dr. Thomas Hutchinson,
Brixham Environmental Laboratory, AstraZeneca, Brixham, Devon, UK.
XI
-------
-------
A. Scope
There has been recent concern for the potential effects of endocrine-disrupting chemicals (EDCs)
on reproduction and development of humans and wildlife species (Colborn et al. 1996). The
Office of Research and Development of the U.S. Environmental Protection Agency (U.S. EPA)
has identified EDC issues as one of six high priority research areas (U.S. EPA 1996a; Kavlock et
al. 1996; Ankley et al. 1997). Further, in response to legislation passed by the U.S. Congress
(Food Quality Protection Act, PL 104-170; Safe Drinking Water Act, PL 104-182), the U.S. EPA
is implementing a screening program for EDCs with specific mechanisms/modes of action
(MOA). To aid in the development of this screening program, the U.S. EPA cosponsored a
series of expert workshops on screening methods (Gray et al 1997; Ankley et al. 1998a; DeVito
et al. 1999), and convened a multi-stakeholder advisory committee (Endocrine Disrupter
Screening and Testing Advisory Committee; EDSTAC) to recommend specific test methods and
screening paradigms for EDCs (U.S. EPA 1998). The focus of these methods is on chemicals
that may affect reproduction and/or development through disruption of physiological processes
controlled by estrogen, androgen, and thyroid hormones. One Tier 1 screening assay
recommended in U.S. EPA (1998) was a short-term (21 d) reproduction test with the fathead
minnow (Pimephales promelas, Rafinesque) designed to identify chemicals that affect processes
controlled by estrogens and androgens. A screening test with fish was considered particularly
important for two reasons: (1) estrogenic/androgenic controls on reproduction/development in
fish may differ significantly enough from that of higher vertebrates such that mammalian (rat)
screening methods may not identify potential EDCs in this important class of animals, and (2) as
-------
opposed to human health effects, there is currently significant evidence of adverse EDC effects in
a variety of wildlife species, including fish (Crisp et al. 1997; Ankley and Giesy 1998). The
purpose of this document is to describe an EDC test method with the fathead minnow designed
to meet the requirements of U.S. EPA (1998).
B. Introduction
The fathead minnow was selected as the test organism for this reproduction assay for a number of
reasons. Attractive attributes of this species include its: (1) widespread geographical distribution,
(2) representation of an ecologically-important family of fish (Cyprinidae), (3) rapid development
and sexual maturation, (4) ease of culturing and testing, (5) common use as a warm-water species
in regulatory testing and decision-making, and (6) the existence of extensive chemical
lexicological databases including information on reproductive physiology and endocrinology that
is valuable in the context of monitoring EDCs.
1. Basic Reproductive Biology ,
The fathead minnow is an omnivorous freshwater fish in the family Cyprinidae. It is a
relatively hardy species with a broad geographic distribution across North America (Devine
1968; Held and Peterka 1974; U.S. EPA 1987). The fathead minnow has a relatively rapid
life-cycle, achieving reproductive maturity within four to five months of hatch (under optimal
conditions). The timing of the reproductive cycle can be controlled effectively through the
2:
-------
use of temperature and photoperiod manipulation (U.S. EPA 1987), thus enabling a lab to
maintain a constant supply of test organisms at a developmental stage suitable for testing. At
maturity, males weigh 4 to 5 g, and females weigh 2 to 3 g. Spawning can successfully occur
at pH values ranging from 6.6 to 9.5 (Mount 1973). Fathead minnows tolerate total alkalinity
concentrations of up to 1,800 mg/L as CaCO3 (McCarraher and Thomas 1968), and turbidity
as high as 15,000 mg/L total solids (Rawson and Moore 1944). The species also is tolerant
of water temperatures ranging from 2 to 33°C (Bardach et al. 1966), and spawns successfully
i
in the temperature range of 15.6 to 29.8°C (Brungs 1971).
Figure 1. Mature male (left) and female (right) fathead minnow, Pimephales promelas.
The adult fathead minnow is sexually dimorphic, with males and females readily
distinguishable from one another when in breeding condition (Fig. 1). As juveniles, the sexes
are similar in appearance. Sexually-mature males develop large nuptial tubercles on the
-------
snout, and an elongated, fleshy, dorsal pad which extends as a narrow band from the nape to
the dorsal fin. The body coloration of males becomes black on the sides except for two wide
light-colored vertical bars (U.S. EPA 1987). In contrast, females generally do not undergo
obvious changes in color or morphology, but do develop a fleshy ovipositor that can be used
to definitively distinguish females from immature/quiescent males (Flickinger 1969). Studies
have demonstrated that secondary sex characteristics in this species are under the control of
sex steroids and, hence, could be affected by chemicals such as estrogen or androgen receptor
agonists or antagonists (Smith 1974; Miles-Richardson et al. 1999a; Harries et al. 2000;
Ankley et al. 2001; Lange et al. 2001). It should be noted, however, that there can be some
degree of ambiguity in differentiating sex even in ostensibly mature fathead minnows. For
example, less dominant males from mass culture situations may resemble females
phenotypically. Alternatively, we also have noted a baseline incidence of reproductively-
active females whose barred coloration patterns can resemble that observed in males. This
seems to occur in situations (such as in the test described herein) where there are multiple
females per test tank; interestingly, in these situations a hierarchal behavior can develop in
which there is a dominant female in the tank which can exhibit male coloration patterns
(MED, unpublished data).
Breeding males are territorial and seek out nest sites which they actively defend against other
males and intruders. Fathead minnows spawn beneath objects (artificial substrates in the lab)
and the buoyant, adhesive eggs stick to each other and to the undersurface of the nesting
substrate. Spawning behavior is characterized by close lateral contact and body vibration
-------
between male and female. After sufficient stimulation, the male presses the female upward,
resulting in the female's urogenital region contacting the substrate, with a concomitant
release of eggs. Milt is released at this time as the pair terminates pressing with an abrupt
separation. This behavior occurs intermittently until an increase in male aggression drives
the female away. The male then guards and tends the nest. The latter activity includes
cleaning the eggs of detritus and agitating the water around the eggs, thereby ventilating them
with oxygenated water. At 25°C, the embryos hatch in about 4 to 5 d (U.S. EPA 1996b).
The number of eggs per spawn can be variable, depending upon the age, size, and condition
of the female (Gale and Buynak 1982). However, the mean number of eggs per spawn under
, ! '
stable laboratory conditions is typically in the range of 50 to 150 (Jensen et al. 2001). Under
the test conditions described herein (25 ± 1°C, 16:8-h light:dark photoperiod), individual
females spawn at intervals of 3 to 4 d (Jensen et al. 2001), creating the potential for a single
female to routinely produce in excess of 500 eggS during a 21-d test.
2. Toxicology
The fathead minnow has been used extensively in short-term (acute) and long-term (partial
life-cycle or complete life-cycle) chemical toxicity studies. They have been widely tested as
a representative warm-water species to provide acute and chronic toxicity data for the
preparation of U.S. EPA national ambient water quality criteria documents. In addition,
fathead minnow toxicity tests with more than 600 chemicals form a unique archival database
(Brooke et al 1984; Geiger et al. 1985; 1986; 1988; 1990; Mayer and Ellersieck 1986; Call
-------
and Geiger 1992) that has been used extensively for a number of purposes, including
quantitative structure-activity relationship (QSAR) modeling (Russom et al. 1991; 1997).
Although the fathead minnow has been tested occasionally in full life-cycle assays
incorporating a variety of reproductive and developmental endpoints (McKim 1977; Lange et
al. 2001), shorter-term tests have been more typical for this species. For example, protocols
are available for 4-d survival and 7-d survival and growth tests that start with either newly
hatched larvae or embryos (U.S. EPA 1993; 1;994). Protocols for slightly longer early life-
stage tests of £28 d are also available; these typically are initiated with embryos, and include
exposure of the embryos, newly hatched fry, and/or juvenile fish to some point prior to sexual
maturation (ASTM 2000b). A comparison ofi short- versus long-term toxicity tests with the
!
fathead minnow (McKim 1977), illustrated the utility of early life-stage tests of 1 to 4 months
in duration in providing an estimate of chemical toxicity over a complete life-cycle.
However, for the purposes of ecological risk assessment, it has been recommended that some
measure of reproduction be incorporated into the partial life-cycle tests (e.g., Suter et al.
1987). None of the above-mentioned short-term protocols address possible effects on
reproduction in adult fish. Therefore, the test protocol described herein includes several!
direct and indirect measures of fecundity as endpoints. This basic test protocol also can
capture endpoints assessed by traditional partial life-cycle and early life-stage protocols
(Ankleye/a/. 2001).
From an ecological perspective, determination of effects of toxicants on reproductive fitness
and, hence, possible population-level impacts clearly is critical (Suter et al. 1987). However,
-------
in terms of screening for chemicals that cause toxicity via MO A of known concern, endpoints
specific to these pathways also are important. Iri recognition of this, endpoints suggested as
desirable for EDC screening in fish models include effects on reproductive behavior,
secondary sex characteristics, gonadosomatic index, gonadal histology, and plasma
concentrations of vitellogenin and sex steroids (p-estradiol, testosterone, 11-ketotestosterone)
(Ahkley et al 1998a; 2001; U.S. EPA 1998; Jensen et al 2001). Recent studies have
assessed the use of these endpoints in EDC studies with the fathead minnow, confirming
their utility in this species, and providing important baseline data in terms of interpretation of
results obtained from these standard test protocols. Specifically, induction of vitellogenin in
response to estrogen receptor agonists (Kramer et al 1998; Panter et al. 1998; 2002; Parks et
al. 1999; Tyler et al 1999; Harries et al 2000; Korte et al 2000; Ankley et al 2001; Lange
et al 2001), and alterations in gonadal histology or secondary sex characteristics associated
with exposure to estrogen or androgen receptor agonists (Smith 1974; Miles-Richardson et
al. 1999a,b; Harries et al 2000; Ankley etdl 2001; Lange et al 2001), have been
characterized in EDC screening studies with the fathead minnow. Finally, there also is an
emerging database concerning the effects of EDCs with known MOA on patterns of
circulating sex steroids in this species (Giesy et al 2000; Makynen et al 2000; Ankley et al.
2001; 2002; Jensen et al 2002).
i
C. Acronyms and Definitions
Acute toxicity - Effects observed in tests of <;96 h in duration.
-------
AQUIRE - AQUatic Information REtrieval system; a U.S. EPA database of aquatic toxicity
information.
ASTER - Assessment Tools for the Evaluation of Risk; a U.S. EPA software program that
combines aquatic toxicity databases and quantitative structure-activity relationship toxicity
prediction models to assist in the development of risk assessments for chemicals.
Chronic toxicity - Effects observed in fathead minnow tests ^28 d in duration.
Dorsal pad - Soft enlargement of flesh on top of the head of sexually-mature male fathead
minnows that extends onto the back of the fish to, or near, the anterior margin of the dorsal
fin.
EDC - Endocrine-disrupting chemical.
ELISA - Enzyme-Linked Immunosorbent Assay; analytical method used for determining
plasma vitellogenin concentration.
p-Estradiol - Major estrogenic sex steroid regulating reproductive function.
Fecundity - Measure of total egg production.
Fertility - Measure(s) of fertilization success as indicated, for example, by actively-dividing
embryonic cells or occurrence of eyed embryos.
GSI - Gonadosomatic Index; gonad weight relative to total body weight
((gonad wt(g)/body wt (g)) x 100).
i.p. injection - Intraperitoneal injection; method of chemical delivery.
11-Ketotestosterone - Major male sex steroid in fish responsible for development of
secondary sex characteristics as well as gonadal development.
LC50 - Concentration lethal to 50% of a group of organisms under specified conditions.
8
-------
LOEC (LOAEC) - Lowest Observed Effect Concentration; lowest concentration of a
chemical that causes a significant adverse effect upon one or more life functions.
MOA - Mechanism/Mode of Action; the mechanism or mode via which a chemical exerts a
toxic response in an organism, ;
NOEC (NOAEC) - No Observed Effect Concentration; highest concentration of a chemical
that does not cause a significant adverse effect upon any life functions.
Nuptial tubercles - Visible external horny outgrowths on the surface of the head of the
sexually-mature male fathead minnow in breeding condition.
Ovipositor - Urogenital structure present in sexually-mature females for egg deposition.
I . , i : i • '' ' •
QSAR - Quantitative Structure-Activity Relationship; a relationship between basic chemical
structure (or property) and a biological response that is described quantitatively.
RIA - RadioImmunoAssay; analytical method used for determining plasma steroid
concentrations.
RPD - Relative Percent Difference; calculation utilized to assess measurement precision.
Saturator - An apparatus capable of generating a saturated stock solution of a chemical that
is relatively insoluble in water.
Subchronic toxicity - Effects observed in fathead minnow tests of >4- and <28 d in duration.
Testosterone - Androgenic sex steroid normally present in both sexes and necessary for
development and maintenance of reproductive function.
Viability - Measure(s) of embryonic development subsequent to fertilization, including
hatching success and normal larval maturation.
-------
Vitellogenin - Precurser to egg yolk protein that occurs normally in the blood of sexually-
mature female fish; it can be induced by estrogen receptor agonists in male fathead minnows.
D. Principle of the Test !
This test is designed as a short-term reproduction assay suitable for identifying chemicals that
affect reproduction or, potentially, development through disruption of any of a number of
pathways, including those controlled by estrogens and/or androgens. Several potentially
sensitive endpoints are assessed. The test is initiated with mature male and female fish that have
a documented history of reproductive success as measured both by fecundity (number of eggs)
and by embryo viability (e.g., hatch). This is established during a pre-exposure phase of 14 to
21 d in the same system/test chambers as will be Utilized for the chemical exposure. During the
subsequent 21-d chemical exposure, survival, reproductive behavior, and secondary sex
characteristics are observed, and fecundity (number of spawns and number of eggs/spawn)
monitored daily. Viability of resultant embryos (e.g., hatching success, developmental rate,
occurrence of malformations) can be assessed in animals held either in clean water or in the same
treatment regime to which the adults were exposed. At conclusion of the 21-d test, blood
samples are collected from the adults for determination of plasma vitellogenin and sex steroids,
and the gonads sampled for measurement of the gbnadosomatic index (GSI) and histological
analyses. Effects in the treatment groups are assessed by comparison to control groups to
determine if any of the endpoints in the exposed fjish are significantly different from those in
10
-------
controls. Those chemicals identified as positive in this test may be advanced for more
comprehensive testing,
£. Information on the Test Chemical
,, ^ i
1. Toxicity to Fish
At least two concentrations of the test chemical should be evaluated in the 21-d reproduction
assay, but additional concentrations are desirable if resources are available (SETAC 1997). It
is also required that the test be conducted as a flow-through, rather than a static-renewal
assay, and that concentrations of the chemical of concern be analytically determined in the
stock solution and test chambers during testing. In addition, it would be desirable to measure
concentrations of the test chemical and relevant metabolites in tissue of the fish at conclusion
of the test. The highest exposure concentration used in the test should be one that is not
lethal to the fish and can be maintained at a near constant level over the duration of the
exposure period. When using a waterborne delivery route this concentration generally should
not exceed water solubility. Depending upon data available for the proposed test chemical,
the following can be used to determine the highest exposure concentration. The order of
priority in terms of utility of available information for establishing this concentration is:
fathead minnow complete life-cycle toxicity data > fathead minnow early life-stage toxicity
data > adult fathead minnow acute toxicity data > chronic toxicity data with surrogate fish
species. Performance of a range-finding toxicity test of 4 to 7 d in length with adult fathead
11
-------
minnows is strongly recommended, particularly if toxicity data are lacking for the fathead
minnow. Fish toxicity data estimated from QS AR models may be used to determine
concentrations to use in a range-finding toxicity test (described below) when no empirical
test data are available for the compound of interest.
To initially ascertain if toxicity data exist for a given test substance a literature search should
be conducted. Possible sources of toxicity information include chemical-specific U.S. EPA
National Ambient Water Quality Criteria or Advisory documents (U.S. EPA, Office of
Water, Washington, DC). These documents describe tests that have generated data of
established quality; they should be consulted to determine if chronic toxicity data for the
chemical(s) of concern exist for the fathead minnow. Another source of acute' and chronic.
toxicity data is the AQUDRE (AQUatic Information REtrieval system) database. This
database is accessible as a component of the larger ECOTOX database (web address:
www.epa.gov/ecotox/). Additional information on accessing this database is available by
e-mail at the following address: ecotox.suppoit@epa.gov.
Data of the highest quality should be used in selecting test concentrations. For example, if
data are available for both flow-through and static tests, the flow-through data generally
should be utilized. Similarly, if data are available for assays in which chemical
concentrations were measured versus tests in Which they were not, the former usually should
be used.
12
-------
If chronic data are available for the fathead minnow, the highest concentration used should be
the LOEC (lowest observable effect concentration), if the adverse effect was upon growth or
reproduction; or the NOEC (no observable effect concentration), if the LOEC was based
upon survival.
If acute, but not chronic, toxicity data for the test chemical are available for the fathead
minnow a divisor may be used to estimate a chronic value from the? 96-h LC50 value. A
divisor of approximately 10 to 12 has been found to be representative of the mean difference
between the 96-h LC50 value and the chronic value (geometric mean of NOEC and LOEC)
for chemicals which have a narcosis MOA (Kenaga 1982; Call et al. 1985). For chemicals
with a more specific MOA, a larger divisor may be appropriate. Selection of a maximum test
concentration from acute toxicity information ideally should be augmented by a range-finding
test, prior to the 21-d reproduction test.
If acute or chronic toxicity data are not available for the proposed test chemical and the
fathead minnow, information should be sought for other fish species. Again, one source of
toxicity data that may prove useful is the AQUIRE database. Such information could be used
to establish a maximum concentration for a rangerfinding test with the fathead minnow. If
only acute toxicity data are available, an appropriate divisor should be applied as described
above. In either case, a range-finding test with the fathead minnow should precede the 21-d
reproduction assay. r
13
-------
If there are no toxicity data available for any fish species, but data exist for chemicals from
the same class which are known or presumed to act by the same toxic MOA as the proposed
test chemical, acute toxicity may be predicted; using a QSAR model. Models for various
classes of chemicals are available, as well as some general rules for model selection
(Bradbury and Lipnick 1990; Veith and Brod&ius 1990; Bradbury 1994; 1995; Russom et al.
1997). Based on this information, a range-finding test should be conducted to help identify
appropriate test concentrations for the 21-d reproduction assay.
Lower test concentrations of the chemical of concern will generally be comprised of a
fractional geometric progression relative to the highest concentration used in the assay; the
magnitude of this fraction will depend somewhat on the number of concentrations utilized.
For example, if only two or three concentrations are used in the test, the lower
concentration(s) should be a factor of 5 to 10 fames less than the highest concentration.
When utilizing a larger number of test chemical concentrations (e.g., ^5), a 50% dilution
series is relatively common, although a logarithmic concentration series is not unusual (e.g.,
100,32,10,3.2, etc.). As a general rule, wheji sensitivity of the analytical chemistry method
for the material of concern is an issue, the lowest concentration tested should be near the
detection limit and higher concentrations should range uniformly to the highest test
concentration. :
-------
2. Physico-Chemical Properties and Chemical Delivery
t _
It is necessary to obtain reliable information concerning the physico-chemical properties of
test chemicals to determine likely behavior in the test system and appropriate exposure
methodology. Factors important in this regard might include water solubility, octanol-water
partition coefficient (K^), melting point, density, volatility, stability in water, and
biodegradability. Information on these characteristics may be obtained from single sources
such as Lewis (1991) or Keith (1997), literature searches, or computerized programs, such as
ASTER (Assessment Tools for the Evaluation of Risk; Russom et al. 1991). ASTER is a
UNIX-based computer program that is not publicly available. To obtain chemical property
information from ASTER, contact should be made via initial e-mail at:
ecbtox.support@epa.gov. Other chemical property databases are commercially available.
For example, the database ENVIROFATE is available through National Technical
Information Service, Springfield, VA (www.ntis.gov); and the databases ECDIN
(Environmental Chemicals Data & Information Network), EFD (Environmental Fate
Database)* LOGKOW, QS AR, and MicroQS AR are available through Technical Database
Services, Inc., New York, NY (www.tds-tds.com).
Knowledge of these physico-chemical properties will allow determination of an appropriate
procedure to use in producing a stock solution of the test chemical, as well as the necessary
rate of renewal (i.e., derived from aqueous stability). For example, if the chemical is highly
soluble in water (^ 1,000 mg/L), stock solutions may be prepared directly in the test water
15
-------
with a slow-stirring method. High-energy stirring, such as with a mechanical blender,
generally should be avoided, as it may produce emulsified droplets of pure compound that
could enter the exposure system. The stock solution may then be pumped directly into a
holding chamber in the test system for subsequent dilution. If the chemical is relatively
insoluble in water, and a liquid at the target test temperature (i.e., melting points <25°C), a
liquid-liquid saturator is recommended for generating stock solutions; whereas, if the
chemical is a solid, a glass wool column saturator is the preferred method of stock solution
generation (Kahl et al. 1999). These methods are described in greater detail in Section G.6.
None of the above methods require solvent or oil carriers. This is preferred because: (1)
certain solvents themselves may result in toxicity and/or undesirable or unexpected
endocrinological responses, (2) testing chemicals above their water solubility (as can
frequently occur through the use of solvents) can result in inaccurate assessments of risk from
the perspective of contaminant bioavailability, and (3) the use of solvents in longer-term tests
can result in a significant degree of "biofilming" associated with microbial activity. If,
however, it is determined that a solvent carrier must be used, several choices are possible
including acetone, dimethylsulfoxide, dimethylformamide, ethanol, ethylene glycol, and
methanol. Toxicity and/or potential endocrinological effects of these solvents have not been
established for the fathead minnow in subchrpnic or chronic assays; however, all are of
relatively low toxicity in 96-h lethality tests (Brooke et al. 1984; Phipps and Holcombe 1985;
Pokier et al. 1986; Geiger et al. 1990; Pillard 1995). If solvent carriers are used, appropriate
solvent controls must be evaluated in addition to non-solvent controls. If it is not possible to
16
-------
administer a chemical via the water, either because of physico-chemical characteristics (low
solubility) or limited chemical availability, it may be necessary to introduce the chemical via
the diet or by intraperitoneal (i.p.) injection (elg., Korte et al. 2000; Kahl et al. 2001). These
routes are less preferable than aqueous administration; this is, in part, due to the fact that
dietary and i.p. routes have not been commonly used with the fathead minnow, which results
in uncertainty in the current identification as to optimized protocols.
3. Range-Finding Toxicity Tests
Given the fact that there will be little acute or chronic toxicity data for many chemicals, or
that existing data are unreliable (e.g., generated from static tests with unmeasured
concentrations), range-finding toxicity tests often will be needed to define appropriate
exposure concentrations for the 21-d reproduction test. The highest concentration for a
range-finding test could be derived from toxicity data for other fish species and/or a QS AR
model (see above). In the absence of any type of empirical or predicted toxicology data, the
range-finding test should start at the solubility limit of the chemical in water. Test
concentrations should decrease by a factor of 10 for each successively lower exposure level.
The range-finding test should ideally be performed with reproductively-mature fathead
minnows under conditions (e.g., fish age, loading rate, temperature, chemical
source/delivery) similar to those to be used in the 21-d reproduction test (described in detail
in Sections G and H, and briefly summarized in Table 1). Ideally, a minimum of two test
chambers (replicates) per concentration, each containing four females and two males, should
-------
be exposed in the range-finding test. Alternative designs of range-finding tests with less
replication could possibly be considered on a case-by-case basis. The exposure period should
be a minimum of 4 to 7 d, with a longer period preferable. The number of mortalities that
occur, and the nature of the concentration (dose)-response curves over time of exposure can
provide critical information in determining the maximum concentration to use in the 21-d
reproduction test. Based on these results, the highest range-finding test concentration that
does not result in increased mortality or signs of overt morbidity (e.g., cessation of feeding),
I
compared to the controls should serve as the highest exposure concentration in the 21-d test.
4. Analytical Determinations
A literature review should be conducted to identify analytical methods that have been used to
measure concentrations of the test substance in water (or, depending on nature of the
exposure route, in food or tissue). An appropriate method should be selected and verified
prior to performance of the range-finding and/or 21-d reproduction test.
F. Validity of the Test
The 21-d reproduction test can be considered valid only if certain conditions are met. These
include: (1) documentation of health of the test animals, as determined from their survival and
reproductive performance during a pre-exposure period, (2) high survival (^90%) of parental
control fish during the exposure period, (3) active; spawning (egg production) of parental control
18
-------
fish, (4) a high rate (;>90%) of fertility in the control group, (5) maintenance of water quality
characteristics (i.e., temperature, dissolved oxygen, pH, alkalinity, total organic carbon, and un-
ionized ammonia) within specified limits (Table 2)* (6) successful analytical measurement of the
test chemical concentrations in the exposure media (e.g., water for aqueous exposures), and (7)
maintenance of chemical exposure concentrations within specified limits (see Section H.2.e for
limits).
G. Description of the Method ;
An overview of the test conditions is provided in Table 1. Specific aspects of the exposure
system(s)/test conduct are presented below.
1. Exposure Apparatus ,
There is no absolute requirement for a particular physical apparatus for the test described
!
herein. Water needs to be delivered to the fish in a consistent manner with controlled
concentrations of the chemical of concern, but several options are available. For example,
the apparatus for flow-through tests may be modified from a proportional diluter system, of
which several designs are available (Mount and Brungs 1967; Defoe 1975; Benoit et al.
1982). Water enters the top of the unit and proceeds through the proportional dilution system
and exposure chambers via gravity feed. Other systems may operate via pumps, as opposed
to gravity feed and be controlled, for example, by electronic units (e.g., DeFoe and Holcombe
19
-------
1997). Water initially enters the system from la common source (same source as the water
used to culture and/or precondition the fish prior to chemical exposure) and is distributed in
equal portions to each replicate control and treatment cell. The water supply must be
adequate to provide a minimum of 960 L (10 L/4 h x four replicates of control, solvent
control, and two chemical treatments) of wateir daily to the test system (approximately
3.5 L/g fish/day). Some testing may require a greater capacity for water delivery (e.g., three
or more chemical concentrations). Further details concerning test systems/chemical delivery
are given below. !
2. Test Chambers
Glass, stainless steel, or other chemically-inert materials should be used for exposure of the
fish to test solutions. Materials with the potential to leach potential endocrine-active
substances such as phthalate esters should be avoided. The dimensions of the test chambers
must be such that reproduction occurs at a consistent rate comparable, for example, to that
achieved under culture conditions. To ensure active spawning and successful fertilization,
four females and two males per replicate exposure chamber are currently recommended. The
minimum recommended chamber size is 40 cjn long, 20 cm wide, and 20 cm high containing
10 L of test solution. Each test chamber should contain three spawning substrates
constructed of stainless steel or PVC pipe (10-20 cm in length) split lengthwise (U.S. EPA
1987). Inflow of water to the chamber should be at the end opposite the outflow to help
ensure a complete replacement with test/clean water.
20
-------
Table 1. Overview of recommended exposure conditions for the fathead minnow 21-d reproduction test.
1. Test type
2. Water temperature
3. Illumination quality
4. Light intensity
5. Photoperiod
6. Test chamber size
7. Test solution volume
8. Volume exchanges of test solutions
9. Flow rate
10. Age of test organisms
11. No. of fish per test chamber
12. No. of treatments
13. No. of replicates per treatment
14. No. of fish per test concentration
15. Feeding regime
16. Aeration
17. Dilution water
18. Chemical dilution factor
19. Chemical exposure duration
20. Primary endpoints
21. Optional endpoints
21. Test acceptability
Flow-through
25 ± 1°C
Fluorescent bulbs (wide spectrum)
10-20 /lE/MP/s, 540-1080 lux, or 50-100 ft-c (ambient
laboratory levels)
16 h light, 8 h dark
1 8 L (40 x 20 x 20 cm) (minimum)
10 L
Minimum of six daily
Approximately 3.5 L/g fish/day
Reproducing adults (120 d minimum)
Four females and two males
Two minimum (plus appropriate controls)
Four minimum ,
Minimum of 16 females and 8 males
Frozen adult brine shrimp twice daily if toxicant exposure
is aqueous or via i.p. injection
None unless dissolved oxygen concentration falls below
4.9mg/L
Clean surface, well, or reconstituted water
5-10 , , '
Adult survival, reproductive behavior, secondary sex
characteristics, gonadosomatic index (GSI) and gonadal
histology, plasma vitellogenin and sex steroid (p-estradiol,
testosterone, 11-ketotestosterone) concentrations,
fecundity, and fertility
Embryo hatch, larval survival, and morphology
Dissolved oxygen ;>60% Of saturation; mean temperature
of 25 ± 16C; 90% survival in the controls; successful egg
production in controls _
21
-------
3. Selection of Test Species
The fathead minnow is used for this test. The rationale for selecting this species has been
described previously which, in summary, is due to their ease of culturing, short life-cycle, and
history of use in toxicity testing. '•
4. Test Water ;
j
Any surface, well, or reconstituted water that results in acceptable survival and reproductive
viability is suitable. The water supply should! be adequate to maintain constant conditions of
water quality during both culture and testing (Table 2). The characteristics of the water
ideally should not impact the availability of the test chemical to the test organisms; thus,
waters high in dissolved or particulate organic carbon should be avoided. Periodic
i
(e.g., yearly) sampling and analysis should be;made of the dilution water for potentially toxic
metals (e.g., Cu++, Ni~, Zn~, Pb+% Hg", Cd+t), major cations (e.g., Ca", Mg**, Na+, K+),
major anions (e.g., Cl", S~), priority pesticides, total organic carbon, suspended solids, and
nitrates.
22
-------
Table 2. Recommended ranges of water quality characteristics for fathead minnow
reproduction.
Water Characteristic
Recommended Range Reference
Temperature (°C)
Dissolved oxygen
(mg/L)
PH
Alkalinity
(mg/L as CaCO3)
Total organic carbon
(mg/L)
Un-ionized ammonia
24.0-26-0
>4.9mg/L
(^60% saturation)
6.5-9.0
I >20
<5
<35
ASTM (2000a,b)
ASTM (2000a,b)
U.S. EPA (1976)
U.S. EPA (1976)
ASTM (2000b)
ASTM (2000a)
5. Test Type « '
The animals must be tested using a flow-through system. The water must be used only once
before discharge to waste. Some test chemicals may require that exposure water be treated
prior to discharge to municipal wastewater treatment plants. For the purposes both of
consistent (parent) chemical exposure and maintenance of adequate water quality (e.g.,
dissolved oxygen, ammonia, etc.), it is desirable to have the test solution reside in the
chamber with the animals for a limited amount of time. The test solution must have a
renewal rate of at least one volume exchange every 4 h, but certain chemicals will need to be
23
-------
renewed more frequently due to their volatility, degradation, or complexation with the test
containers.
6. Stock Solutions (Aqueous Exposures)
I
Chemical stock solutions can be generated using a number of approaches, and delivered via
either proportional dilution systems or stock pumping systems (Mount and Brungs 1967;
DeFoe 1975; Benoit et al. 1982; DeFoe and Rolcombe 1997). The method of choice for
preparation of stock solution(s) must provide for an adequate volume of water to achieve
desired test concentrations for a flow-through test. It is preferable that stock solutions be
prepared in water without the use of solvents or solubility enhancers. As described below,
different approaches have been used to successfully achieve appropriate test concentrations
for chemicals of various physico-chemical characteristics in the absence of solvents. In some
instances, the use of a solvent for chemical delivery in waterborne exposures will be
unavoidable. •
a. Solid-Liquid (slow-stir) Saturator. A ^olid-liquid saturator is designed to use the
surface of the inside of a glass carboy ior vessel as an area which, when coated with a
test chemical, generates a stock solution at or near saturation. This is a slight
adaptation of a system in which stock [solutions are prepared directly by dissolving
test chemical in water. Compounds which are solids and moderately soluble in water
are best suited for this method. Generally, these types of chemicals are not soluble
24
-------
enough to mix in stock vessels for immediate use, but not so insoluble that it is
necessary to coat substrates with a large surface area, such as a glass wool column
saturation unit described below. To determine the effective use of this method in a
test system, analytical measurements should be taken from the chemically-coated
vessel at a range of flow rates over time.
.' ' • i •
The test chemical is dissolved in a solvent (e.g., acetone, ethanol) before application
on the glass vessel (> 1L), with the amount of solvent varying due to vessel size and
amount of chemical. Coating of the glass with the chemical requires an even
distribution, which can be successfully achieved by rolling the container under a low
flow of air or nitrogen gas to evaporate solvent from the system. Vessels that can be
placed on a rolling mill simplify coating the chemical evenly. Upon solvent dryness,
usually identified by a crystalline design, the vessel is filled with test water to the
desired level, fitted with the stopper assembly and secured. It is important that the ai
space in the saturator be completely sealed from air exit or entry during use of the
saturator, and that it be possible to monitor the rate of incoming make-up water at the
rate of the diluent being removed. Water input occurs just above the water level in
the vessel, with an output tube located near the bottom. The diameter of the output
tubing is best when matched with that used on the dispensing pump. Rapid stirring
within the vessel is produced using a magnetic stir bar and stir motor. The presence
of a vortex at the air/liquid interface generally is sufficient for adequate stock mixing.
This type of saturator has been used successfully with the model antiandrogen
air
25-
-------
flutamide (Jensen et al. 2002). In that study, a 3.5 g quantity of chemical was
dissolved in 40 ml of acetone. With an 8 ml/min flow rate through an 18 L carboy,
stock concentrations were generated at about 21 mg/L for up to 7 d. Therefore, three
coated vessels supplied a usable stock! for the 21 d of chemical exposure.
b. Liquid-Liquid (slow-stir) Saturator. In a liquid-liquid system, water that is nearly
saturated with the test compound is pumped from the saturator vessel to the diluter
system, and at the same time replacement water is added back to the saturator at the
rate that it was withdrawn. This type 6f saturator has been successfully used with
chemicals that are liquid at room temperature with water solubilities between 3.2 and
20,000 mg/L (Kahl et al. 1999). A glass container of sufficient volume is required to
achieve a surface to volume ratio between chemical and water that is adequate to
generate a volume of test solution near water saturation for the chemical. The system
generally consists of a glass container^ 1 L) containing water, test chemical, a
magnetic stir bar, and air space (Fig. 2). The container should have a narrow neck
into which a neoprene stopper can fit.! Four stainless steel or glass tubes are inserted
through the stopper to permit input of [chemical, venting, input of dilution water, and
output of the test solution. If the test chemical has a specific gravity <1, it will float
on the surface of the water and the dilution water input tube should be positioned
above the solution level, and the output tube near the bottom of the container
(Fig. 2A). If the specific gravity of the test chemical is >1, the chemical will be close
i
to the bottom of the container. In this'case, the dilution water input tube should be
26
-------
near the bottom, and the output tube in the upper half of the solution (Fig. 2B). The
container is placed on a magnetic stir bar drive motor, and the stir bar rotated rapidly
enough to create a vortex at the surface of the solution, but not so vigorously as to
cause droplets of the test chemical to separate from the pure chemical mass into the
water portion of the partitioned solutions. It is important that the air space in the
saturator be completely sealed from air exit or entry during use of the saturator. This
air space will come into equilibrium with the water, and will not cause differences in
solution concentrations with time. A pump is attached either to the dilution water
input tube or the output tube which delivers test solution to the test system. Liquid-
liquid saturators have been used successfully in exposures of aquatic organisms to a
variety of potential EDCs, including di-n-butyls butylbenzyl, and di(2)-ethylhexyl
phthalate (Call et al. 2001), 1,4-dinitrobenzene (Geiger et al. 1985), methoprene
(Ankley et al. 1998b), and nonylphenol (Kahl et al. 1997).
A. Specific Gravity <1
output
input-
B. Specific Gravity >1
• output
I
input
air,
aqueous
aqueous
organi
Figure 2. Liquid-liquid (slow-stir) saturation unit.
.27,
-------
c. Glass Wool Column Saturator. This type of saturator has been used successfully with
[
chemicals having water solubilities between 0.0074 and 104 mg/L (Kahl et al. 1999).
Glass wool column saturators have be£n used successfully to generate stable water
concentrations near solubility with chemicals that are liquids, but have specific
gravities near 1, and do not maintain a discrete layer separate from the dilution water
when mixed with a stir bar. This type i of saturator can also be used to generate test
concentrations near water solubility with chemicals that are solids at room
temperature. The system is comprised of a clear glass tube approximately 2.5 cm in
I
diameter and 1 m in length bent in a U-shape (primarily for economy of space) and
packed with cleaned (solvent-rinsed) glass wool (Fig. 3). The test chemical is
dissolved in a solvent (e.g., acetone, ethanol) which is added to the glass wool and
drawn through the glass wool with a vjacuum applied at one end. About one-half of
the chemical solution should be added from each end of the glass tube to ensure
uniform coating of the glass wool (for greater detail, see Kahl et al, 1999). When the
column is completely vacated of the solvent, both ends of the tube are plugged with
stoppers. Each stopper has a tube inserted and a pump is used to push water through
the column. For some chemicals, a single pass of water through the column may
i
result in the desired chemical concentration in the solution. Other, chemicals may
1 ' !
need several passes utilizing a recirculation system that pumps to a stock reservoir
and back through the column. In the Ijatter case, the concentrated stock is drawn from
the stock reservoir. This type of stock solution-generating system has been used
successfully with a variety of potential EDCs including DDT, ODD, and DDE (Hoke
i
I
28
-------
et al 1994), dieldrin (Hoke et al. 1995), methoxychlor and methyltestosterone
(Ankley et al: 2001), polycyclic aromatic hydrocarbons (Erickson et al. 1999), and
vinclozolin (Makynen et al. 2000).
The capacity of any of these saturators to maintain a constant concentration of
chemical in solution will diminish with time of use. A general rule is to use at least
twice the amount of chemical as the test may require to maintain a uniform
concentration over the entire exposure period. A determination of the proper amount
to "load" into the saturator can be calculated from information on water solubility of
the test chemical (either from the literature or preliminary testing), volume of
solutions in test chambers, number of dilutions, number of replicates per treatment,
number of solution volume exchanges daily, and duration of the test. Multiple
columns may be required over the course of one experiment.
Figure 3. Glass wool column saturation unit.
29
-------
d. Solvent Carriers. Certain chemicals m^y be insufficiently soluble in water or too
j
unstable to use a saturator and generate a solution of sufficient concentration to
conduct a test. It may be necessary to Dissolve this type of chemical in a solvent
carrier to achieve a test concentration near water solubility. Solvents may be injected
into a dilution system for proportional {dilution or directly injected into dilution water
to prepare test solutions of desired chemical concentrations. In these experiments
corresponding solvent-only control chambers are required.
7. Number of Treatments
The minimum number of treatments (test concentrations) is two (three if a concentration
(dose)-response is sought), in addition to appropriate controls. Additional test concentrations
are highly desirable but are, of course, dependent upon availability of resources. A single
control treatment is required in the case of aqueous exposures. Two control treatments are
required if chemical administration is via either solvent carrier, i.p. injection, or feeding (i.e.,
a solvent or sham-treated control in addition to a non-solvent or non-injected control).
8. Replication •
The test chamber is considered to be the experimental unit or replicate. As the number of
i
replicates per treatment increases, the number of degrees of freedom increases, the width of
the confidence interval for point estimates decireases and, as a consequence, power of the
301
-------
hypothesis test increases. The sensitivity of the test will be dependent upon the inherent
variability of the endpoint under consideration as a function of the number of replicates
employed in the treatments and control. Thus, every effort should be made to reduce among-
sample variability of the measured endpoints. The greater the variance in measured effects
p : I- •
between individuals within a replicate and the resultant means within a treatment, the greater
the number of replicates needed to measure'significance at any confidence level.
It is important that interpretation of the data facilitates evaluation of the potential for Type I
and Type n errors. Type I error is the error of rejecting a null hypothesis (no difference
between the control and treatment means) that is true; Type n error is the error of failing to
reject the null hypothesis when it is false. For the purposes of a screening test, both types of
error are critical, but the Type n error is more important because it could result in not
detecting a possible EDC. Avoidance of the Type n error is accomplished by increasing the
power (1-P) of the test. Many statistical tests with biological data are conducted with an a
value of 0.05 and a p value of 0.20, where the probability of a Type I error is 5% and a Type
n error is 20%. A more conservative Type H error probability of 10% (p = 0.10) may be
desirable to increase sensitivity of the test. Type n error minimization (i.e., increased test
power) requires either, that the test be conducted with a greater number of replicates per
treatment or that variability between treatment measurements be reduced (Fig. 4).
31
-------
1.2
3 I 0.6
ii
E 9
sl 0.4
H
0.0
5 10 . 15 20 25
Number of Replicates Required
30
Figure 4. Number of replicates required per treatment to detect differences between
treatment means with ANOVA using three treatments when a = 0.05 and p = 0.10 or
0.20. Ratios are the desired minimum detectable difference in treatment means divided
by the pooled standard deviation.
It is convenient to express the ratio of the difference between control/treatment means
relative to their pooled standard deviation to demonstrate the impact of sample variability and
sample size on detecting minimum differences (Fig. 4). For example, to demonstrate
differences at the 95% confidence level (a = Q.05) and with a power of 0.80 with four
replicates per treatment, the variance between! the means of the treatment(s) and control must
be less than 50% of the difference between th0 means for which a significant difference may
exist. If the variance is greater than 50%, more replicates will be required to detect
differences. In the test report, power of the statistical analysis together with the confidence
value should be reported.
32
-------
For the purposes of this reproduction test, it is currently recommended that a minimum of
four replicates (test chambers) be used per treatment. As with the number of concentrations
tested, additional replication is desirable depending upon available resources. Further, as
discussed above, depending upon established endpoint variability, replication may be
increased in a study-specific manner. Unfortunately, because this test has not been used
routinely, estimates of endpoint-specific variation are not yet robust enough for routine
a priori design of tests with known power.
'!
9. Performance Standard Test
A suggested approach for assessing the ability of a laboratory to conduct this assay is
demonstration of expected test results with one or more known EDC(s). Conducting this
i' i
type of test should reduce the frequency of the Type n error (i.e., having no effect when there
is an expected effect) by demonstrating that the test system is performing reliably from both
qualitative and quantitative (statistical) perspectives. Thus, it is recommended that each
laboratory planning to perform this reproduction assay conduct a performance standard test
with one or more known EDC(s). The results of this test should be reviewed and approved
by knowledgeable scientific staff from outside of the laboratory prior to the initiation of
routine screening assays. Chemicals potentially useful for this type of test might include
strong estrogen receptor agonists such as p-estradiol or 17cc-ethynylestradiol; or an aromatase
inhibitor such as fadrozole (Kramer et al 1998; Panter et al. 1998; 2002; Miles-Richardson
et al. 1999a; Tyler et al. 1999; Korte et al. 2000; Ankley et al. 2002).
33 ''.' .
-------
10. Reference Toxicant Tests
As a routine quality assurance option, it is desirable that reference toxicant tests be performed
|
regularly (e.g., once annually), or concurrently with approximately 10% of other tests, to help
verify aspects of general health of the test organism. This assay can be conducted using
chemicals that are not necessarily EDCs, and pan be comprised of short-term tests (e.g., 96-h)
i
rather than the 21-d reproduction test. Some suggested reference toxicants include sodium
chloride, potassium chloride, cadmium chloride, copper sulfate, sodium dodecyl sulfate, and
potassium dichromate. If a reference test doejs not generate expected results, the test should
be repeated immediately. Results from reproduction tests conducted during the interval of
time since the last successful reference test arje suspect. Reference toxicity tests must be
conducted repeatedly with the same reference^ toxicant, at the same concentrations, in the
i
same dilution water, using the same data analysis methods. A quality control chart is used to
!
track the test results over time. The control chart should demonstrate that the test results are
within ±2 standard deviations of the central tendency once several tests have been performed
to establish the central tendency (Lewis et al. \ 1994). Reference toxicant tests could prove to
be especially useful in situations where test animals are obtained in different lots from
i
commercial vendors. !
34
-------
H. Procedure
1. Pre-Exposure Reproduction
i i ' ' ,
a. Selection of Fish. The pre-exposure phase of the test should be started with animals
that have achieved reproductive maturity (typically a minimum of 120 to 180 d old),
as evidenced by initial development of secondary sex characteristics, but have not
been held in a culture/test situation conducive to routine spawning. Ideally, these fish
would be the offspring of several pairs of adults. Four females and two males should
be randomly assigned to the replicate test chambers at each anticipated treatment
concentration. Identification of gender may be difficult to resolve for some fish; these
animals should not be used for the test. At this stage in development, males will
exhibit nuptial tubercles, while females possess an ovipositor; in addition, males tend
to be larger and darker than females from the same cohort (U.S. EPA 1987; Jensen et
al. 2001). Some experience in working with this species will be required to
accurately identify sex based on phenotypic characteristics, in particular, just prior to
active spawning.
In addition to the number of replicates (e.g., four) which actually will be exposed at
each concentration, additional test chambers should be started; these can serve as
"replacement" units for chambers where fish sex was misidentified, pre-exposure
35
-------
spawning does not occur, and/or mortality is observed during the pre-exposure
observation period.
b. Conditions. Pre-exposure reproduction observations must be performed using the
same regime as described in Sections (3.1, G.2, and H.2. The pre-exposure initiation
i
of sexual reproduction must occur in the same chambers and under the same
conditions (i.e., water characteristics, flow rate, lighting, temperature, feeding, and
maintenance) as the 21-d reproduction! test, to prevent unnecessary agitation of the
fish at initiation of exposure to the test chemical. This monitoring phase establishes
both the reproductive success of the spawning animals, and provides a quantitative
chamber-specific baseline for potential statistical comparison after initiation of the
chemical exposure. That is, each test Chamber can serve as its own control.
!
c. Observations. Organisms in each test (chamber should be observed daily for obvious
i
alterations in secondary sex characteristics (i.e., nuptial tubercles, dorsal pad, and
darkened coloration in males; and distended abdomen and swollen ovipositor in
females), and reproductive behavior. '
Observations should be made daily of ispawning activity, and records maintained for
the number of spawns and number of bggs per spawn. It also would be desirable to
assess aspects of development of resultant embryos, such as fertility, embryo hatch,
and gross appearance of newly-hatched larvae. If the animals are not reproducing
36
-------
satisfactorily by 21 d, the pre-exposure portion should be terminated, and a thorough
evaluation performed of the test system and fish to determine why satisfactory
reproduction had not occurred.
>
d. Suitability for Testing. The pre-exposure period may last from 14 (minimum) to 21
(maximum) d. The exposure can be initiated once it has been established that regular
spawning is occurring in each test chamber every 3 to 4 d. Based on information
from tests conducted at MED to date (Ankley etal 2001; 2002; Jensen et al. 2001;
2002), a minimal fecundity of 15 eggs/f/d would be expected for the test conditions
described herein. The eggs should be successfully fertilized by the males, with a
minimal fertility of 90%. Embryo hatch should be 2:90% and fry must be normal in
appearance and swimming behavior; this is particularly critical when assessment of
the F! generation is of interest. If these criteria of fecundity, fertility, embryo hatch,
and fry appearance are not met, fish in the test chambers that fail to meet the criteria
should not be used. If the criteria cari not be met in all four of the test chambers
during the pre-test conditioning, the fish should be discarded, the test system cleaned
and evaluated, any necessary changes made, and a pre-exposure period initiated with
a new group of fish.
37
-------
2. Conditions of the 21-d Reproduction Tes
a. Duration. The target exposure duration is 21 d. This exposure period should be
sufficient for healthy females to produce several clutches of eggs. As such, this time
period allows for a robust assessment bf reproductive behavior, fecundity, fertility,
and, if desired, embryo development and hatching success. In addition, the 21-d test
i
period should help optimize exposure [of the fish to relatively hydrophobic chemicals
i
that require a period of time to reach steady-state concentrations in the animal, m the
I
context of screening, instances may occur in which a test is terminated in less than
21 d. For example, it may be that the test chemical is clearly causing such significant
adverse effects (e.g., lethality, cessation of spawning, observation of abnormal
i
secondary sex characteristics, or behavior not consistent with reproduction, etc.) that
i-
i
it is clear that: (1) the test should be reinitiated at lower chemical concentrations, or
(2) from a screening perspective, an effect consistent with an EDC has been
established.
b. Loading. The total mass of fish in thk test chambers must not affect the results of the
i
test through alterations in water quality. Specifically, the total mass must be low
enough to: (1) allow a ;>60% or greater dissolved oxygen saturation in the test
solutions at all times (i.e., :>4.9 mg/L {at 25°C), (2) prevent concentrations of
i
metabolic products (predominantly ammonia) from exceeding levels that affect
animal behavior and health, and (3) prevent stress due to crowding. It is
I
38
-------
recommended for a test of this type that the loading not exceed 0.5 g fish/L of
solution passing through the test chamber per 24 h, or an instantaneous loading rate of
5 g fish/L (ASTM 2000a). The instantaneous loading rate for the test conditions
suggested herein (i.e., four females and two males/10 L) is approximately 2 g fish/L.
c. Light and Temperature. Room light intensity of 100-foot candles during the lighted
portion of the day is recommended. The light can be provided by fluorescent or
incandescent lights. A photoperiod of 16 h light and 8 h dark, at a temperature of
25 ± 1°C is recommended. During a test, the water temperature should be monitored
continuously in at least one test chamber, and daily in all others, and documented as
being maintained within ±1°C.
d. Feeding. In tests where the route of test chemical administration is via either aqueous
exposure or i.p. injection, the food should be frozen brine shrimp (Artemia) which
have been thawed. The test animals should be fed the brine shrimp ad libitum twice
daily. The frozen brine shrimp should be obtained from an established commercial
source, and should have been demonstrated to result in acceptable survival, growth,
and reproduction of the fathead minnow under culture conditions. Chemical
characterization of the shrimp for common environmental contaminants that might
affect endocrine function (e.g., organochlorine pesticides, PCBs) would be prudent.
The amount of brine shrimp fed to each test chamber should be enough to nearly
39
-------
satiate the fish with little, if any, food Remaining at the bottom of the test chamber at
the end of each feeding. \
I
If test chemical administration is via the diet, the fish should be fed commercially
available fish food with the chemical incorporated into it. One method of adding the
chemical of concern to the food employs a rolling technique, in which a glass vessel
containing the food and the chemical dissolved in an appropriate solvent (e.g.,
acetone, hexane) is mixed on a roller mill (Fernandez et al. 1998). Large lots of food
can be made in a few minutes and, if storage life permits, refrigerated for use during
the study. The chemical in the food should be measured for uniformity and stability
during storage. i
i
It should be noted that there has been little, if any, lexicological research in which
fathead minnows have been dosed visi incorporation of chemical into the diet. And,
i
there is some uncertainty as to palatabjility/suitability of commercial pelleted feed to
i
fathead minnows, particularly in longer-term (reproduction) studies. For these
i 11 .
reasons, the dietary route, although conceptually reasonable is, at present, the least
i
preferred manner of test chemical delivery.
e. Test Concentrations. The highest concentration used for the 21-d reproduction test
should not have caused significant mortality in the range-finding assay (note that this
[
may be at water solubility), and the lojwer concentration should normally be at a factor
40
-------
of five to 10 times lower than the highest test concentration. The use of two test
concentrations in this fashion not only enables at least some consideration of
unexpected concentration (dose)-response relationships, but also provides a safety
factor (lower concentration) should the high concentration prove to be lethal over the
21-d exposure. Alternatively, some studies may be undertaken using £ three test
concentrations using a semi-log concentration series of 100,32,10,3.2 and so forth
(Panter et al. 1998). Concentrations of test chemical at each treatment level must be
sufficiently uniform during the test to ensure that the fish are exposed to desired
concentrations. In an aqueous exposure, concentrations at different levels should riot
overlap (e.g., due to cyclical trends associated with degradation). At a given exposure
level, the measured concentration of the test chemical should not be less than 50% of
the time-weighted mean concentration for more than 10% of the duration of the test
(ASTM 2000b). In addition, the measured concentration should not be greater than
30% higher than the mean concentration for more than 5% of the duration of the test
(ASTM 2000b). It should be recognized that once the exposure starts, there generally
will be a gradual increase in the concentration during the first day. For example, with
the recommended exchange rate of six volumes/d (42 ml/min) and a volume of 10 L
per chamber, 95% replacement will not occur until approximately 12 h after test
inititation (Sprague 1969). Analytical determinations of chemical concentrations
should be made as frequently as possible with a minimum frequency of once weekly
in at least one of the replicate test chambers at each concentration, and preferably in
all chambers on the first and last days of the test. The variability of both the sampling
41 .
-------
and the analytical procedures should be
that observed variability in
techniques.
determined before initiating a test to ensure
measured concentrations is not due solely to analytical
If the test chemical exposure route is by i.p. injection or the diet, tissue residue
analyses should demonstrate a direct relationship between body burden and treatment.
Further, it would be desirable to analyjze the stock solution used either to spike the
j
diet or for the i.p. injections to verify chemical concentrations prior to treatment of
the animals. !
I
The test chemical must be as pure as practical (>90% minimally, unless the chemical
does not exist in this purity range).
f. Controls. Dependent upon route of chemical administration, one or more sets of
controls will be required. If the chemical exposure route is aqueous, only a non-
exposed control is required. If the chemical exposure route is aqueous, but with a
solvent carrier, a solvent control is required in addition to a control that is not
exposed to solvent. If the route of exposure is via i.p. injection, a sham- (e.g., corn
oil) injected control is required in addition to a non-injected control. If the route of
chemical exposure is via the diet, controls both without and with manipulated (i.e.,
spiked with solvent) food are required.
42
-------
g. Fish. Organisms for the initiation of a laboratory culture should be obtained from a
source which has a verified fathead minnow culture. Brood stock must be adapted to
laboratory conditions and free of disease. Embryos make the best initial stock offish
because they are easiest to transport and most likely to be free of disease. Purchased
fish need to be observed in the testing laboratory for a time suitable to ensure that
they are in good health. Stock from wild populations should be avoided unless
cultured through at least one generation to ensure they are disease-free and of
adequate vigor (U.S. EPA 1987). Animals for starter cultures are available from the
U.S. EPA Aquatic Biology Branch, Quality Assurance Research Division, EMSL-
Cincinnati Newtown Facility, 3411 Church Street, Newtown, OH 45244 (Telephone:
513-533-8114); the U.S. EPA Mid-Continent Ecology Division, 6201 Congdon
Boulevard, Duluth, MN 55804 (Telephone: 218-529-5000); or any of several
commercial suppliers with species-verified stocks^ such as Environmental Consulting
and Testing, 1423 North Eighth Street, Superior, WI54880 (Telephone: 715-392-
6635 or 800-377-3657; web address: www.ectesting.com); or Aquatic Research
Organisms, Hampton, NH 03842 (Telephone: 603-926-1650; e-mail:
arofish@aol.com; web address: www.holidayjunction.com/aro).
For laboratories not intending to start a fathead minnow culture, animals of a known
age (prior to maturity) may be obtained from commercial suppliers for testing. If this
is done, animals should be held for at least 1 month prior to initiating the pre-
exposure period for the test.
43
-------
h. Cleaning of Test Chambers. All components of the test apparatus that will have
i
contact with water or chemical must have been adequately cleaned with detergent,
solvents, and deionized water to prevent the occurrence of disease organisms or
chemical residues from any previous vise of the apparatus. During the pre-exposure
period and the test, chambers need regular cleaning to remove biological debris
(excess food, excrement). This can b6 effectively accomplished with a small (0.8-
1.0 cm i.d.) Tygon® siphon tube. Aftjer one week, algal and fungal growth may begin
to appear, interfering with observations of fish, reducing the adherence of eggs to the
spawning substrates, and potentially reducing bioavailability of the test chemical.
The spawning substrates should be removed from the test chamber and scrubbed.
The growth on the inside of the test chamber can be removed by carefully (to avoid
agitating the fish) scraping the inside walls with a sponge or razor blade inserted into
a long-handled holder. Once the material scraped from the chamber walls has settled
to the bottom, it can be removed with
a siphon tube, and the spawning substrates
replaced. It is essential to treat all test and control chambers identically to avoid
introducing bias.
3. Test Chemical Administration
a. Flow-Through Aqueous-Delivery.
achieved by blending known
44
Test concentrations of the chemical can be
concentrations of stock solution with dilution water in
-------
mixing cells of the exposure apparatus, from which solutions of desired
concentrations are allocated to test chambers. To achieve comparable flow rates of
water among the treatment groups, separate reservoirs containing different
concentrations of stock solution should be prepared for each treatment level, and the
pumping rate from each reservoir should be equivalent. The stock solutions may be
prepared as specified previously (Section G.6), depending upon characteristics of the
test chemical. Flow rates to the exposure chambers should deliver a minimum of six
volume exchanges daily.
b. Flow-Through Aqueous/Solvent-Delivery. For tests in which a solvent or dispersant
is used to facilitate delivery of the test chemical, a solvent or dispersant control must
be included in the experimental design. This control should be identical in all aspects
to the non-solvent control, except that it should contain the solvent or dispersant at
the maximum concentration used in the different test chemical treatment levels. The
delivery system should be maintained as described above for the water-delivery
system. It may be necessary to increase the frequency of cleaning the delivery system
and test chambers, due to the promotion of microbial growth by the solvent. Flow
rates to the exposure chambers should deliver a minimum of six volume exchanges
daily.
c. Intraperitoneal (i.p.) Injection. Certain chemicals may be so insoluble in water that
aqueous exposures are not feasible. Alternatively, availability of the test material
1 i , '
45
-------
i •
may be so limited that a flow-through jaqueous exposure is not feasible. In such
cases, exposure via the diet or by i.p. injection may be required. When the route of
chemical administration is via i.p. inje'ction, two types of controls must be employed,
a non-injected control and a sham-injected control. The test chemical should be
dissolved in a non-toxic carrier, such as corn oil or a mixture of ethanol/corn oil
(Korte et al 2000; Kahl et al 2001). ;The injection procedure involves first lightly
anaesthetizing the fish with MS-222 (koo mg/L, buffered with 200 mg NaHCO3/L),
then carefully holding the fish in one Ijand in an inverted position, and inserting a
sterile needle containing the desired vblume (and concentration) of test solution just
through the abdominal body wall (Figi 5). The maximum needle size should be
26 gauge (ca. 1 cm in length), and no more than 10 ul of solution/g body weight
should be injected in an individual fish. The needle should be inserted just anterior to
the vent with the needle tip oriented ahteriorly and at an angle of approximately 30 to
I
45° from the horizontal position. The' orifice at the tip of the needle must be
completely inserted through the body Wall without entering internal organs to ensure
i
that the entire volume desired is injected into the fish. After injection, the needle tip
is withdrawn and the point of injectiofi carefully examined to determine that there has
not been any loss of solution. The injjection process should be accomplished quickly
(e.g., within 30 s), after which the fish should be immediately returned to its
|
respective test chamber. Injected fishj should be observed over the next several hours
to ensure that they have recovered from the anaesthetic and have not suffered any ill
effects from the injection. Depending upon the kinetics of metabolism/depuratii
i
i
ion of
46
-------
the test material, multiple injections may be required during the 21-d test. Data from
the MED laboratory indicate that i.p. injections at a frequency of up to one/week
should not affect the reproductive and endocrinological endpoints associated with the
test(Kahle/a/. 2001).
When using the i.p. exposure route, flow rates (of clean water) to exposure chambers
should deliver a minimum of six volume exchanges daily.
Figure 5. Handling technique for administering test chemicals to fathead
minnows by i.p. injection.
d. Dietary Exposure. In cases where a dietary route of chemical administration is used,
commercial fish feeds are recommended. One procedure that has been described for
preparing a diet containing a highly insoluble chemical (2,3,7,8-tetrachlorodibenzo-p-
dioxin) utilized an apparatus that delivered the test chemical dissolved in n-hexane to
47
-------
the food pellets in a modified glass jar| on a roller mill (Fernandez et al. 1998). The
spiked hexane is removed from the food using a vacuum, and condensed and
I
collected in a cold trap. It is recommended that a procedure similar to this be used to
minimize exposure of the fish to the cjarrier solvent and, at the same time, achieve an
acceptable degree of homogeneity of test chemical in the food. Two types of controls
I
are needed when conducting dietary exposures. One should receive equal amounts of
|
food that has not received any treatment of either solvent carrier or test chemical, and
the second control should receive foo
-------
4. Analytical Determinations and Measurements
a. Test Chemical. For tests with an aqueous route of administration, concentration of
the stock solution must be measured initially and at least once weekly. In addition,
the test chemical should be measured in the exposure water at least once weekly
during the 21-d exposure. In the range-finding tests, the behavior of the chemical in
solution also should be determined; this is a good opportunity to assess, under
simulated test conditions, performance of the stock solution generation/delivery
system. For chemicals that do not require range-finding tests, performance of the
exposure system should be evaluated independently prior to testing. During the 21-d
reproduction test, measurements should be made in at least one replicate test chamber
of all exposure concentrations on each sampling day. The analytical procedures
should follow U.S. EPA-approved or other validated standard methods, if available.
When appropriate, reagent blanks, recoveries, and standards should be included
whenever samples are analyzed. Balances and analytical instruments should have
properly updated records of calibration or service, as appropriate. Standard reference
materials should be analyzed, when available, to document accuracy of analyses.
Recovery of the test chemical from the exposure medium should be determined from
a sufficient number of samples to allow an assessment of accuracy of the analytical
process. A minimum of three samples is required; however, six or more analyses are
preferred. Precision of the analyses is determined by collecting duplicate samples at a
49
-------
frequency of at least 10% of the total samples analyzed. A relative percent difference
i
(RPD) is calculated for each set of duplicate analyses, and an overall mean and
standard deviation calculated for the en'tire number of duplicates analyzed. The
i
method detection and quantitation limits for test chemicals must be defined.
b. Water Characteristics. Water quality characteristics must be measured regularly to
document that they were sufficiently uniform to ensure that adverse biological effects
due to water quality did not affect the outcome of the test. Temperature must be
monitored daily and maintained at 25 ± 1°C. Dissolved oxygen in the test solutions
must also be monitored daily to ensure ;a minimum concentration of 4.9 mg/L (60%
saturation at 25°C). Alkalinity, hardness, pH, and un-ionized ammonia, as a
minimum, should be measured at least weekly during the test in a high and low test
concentration. In addition, since particjulates and organic carbon may influence the
bioavailability of many test substances,! concentrations of both total organic carbon
i
and particulate matter should be determined and documented as ^5.0 mg/L (ASTM
2000b). Table 2 summarizes recommended ranges of values for water quality
characteristics. Sample data forms, such as the example provided (Appendix A,
Form A-l), should be prepared in advance and completed on a daily basis.
50
-------
5. Measurement of Test Endpoints
Observations of the test animals (described below) should be made daily and/or at conclusion
of the test.
a. Survival of Adults. Daily assessment of survival is necessary to provide a basis for
expression and interpretation of reproductive output, that is, number of eggs per
female per day. Unless unacceptable water quality excursions and/or disease occur, it
is rare to observe mortality in untreated control animals. In animals exposed to the
test chemical, overt lethality may occur, particularly in later portions of the assay not
reflective of the initial (shorter) range-finding test. Mortality should be recorded daily
on a form such as the example provided (Appendix A, Form A-l),
b. Behavior of Adults. Any abnormal behavior (relative to control animals) observed
during the course of the chemical exposure should be monitored. This might include
signs of general toxicity, including hyperventilation, uncoordinated swimming,
atypical quiescence, loss of equilibrium, or abnormal feeding. From the standpoint of
FJDC screening, alterations in reproductive behavior, particularly loss of territorial
aggressiveness by males, may be affected by FJDCs (Kime 1998). Detailed
descriptions of normal behaviors and the abnormal behaviors should be recorded to
compare the changes that may be observed during the test. Because of the relative
51
-------
subjectivity of this endpoint, it may be jnecessary to document behavioral alterations
via photograph or videotape. !
i
i
i
c. Fecundity. Fecundity, or total egg production, should be determined on the basis of
i
surviving females per reproductive day|for each test chamber. For example, if all four
i
females survived the treatment in a given chamber in a 21-d exposure, there would be
84 female reproductive days. If one or more females dies during the course of the
exposure, the number of female reproductive days would be reduced accordingly in a
j
"pro-rated" fashion. j
i
[
! I
Fathead minnows usually spawn in the early morning hours so they should not be
i
disturbed during this time except for a morning feeding. The spawning substrates
should be checked for the presence of eggs in the late morning. This period of time
allows the opportunity for spawning to
occur and the eggs to water harden. It is
necessary to remove the substrates from the test chambers to prevent the eggs from
being eaten by the adults, and for purposes of determining egg count and fertility. If
no eggs are present, the spawning substrate is left in the test chamber; new substrates
should be added to replace any that were removed. Data forms to record fecundity,
i
such as the example provided (Appendix A, Form A-2), should be prepared in
|
advance and completed on a daily basis.
52
-------
d. Fertility. After the spawning substrate has been removed from the test chamber, the
eggs should be carefully rolled off the substrate with a gentle circular motion of an
index finger (Gast and Brungs 1973) and visually inspected under appropriate
magnification. If more than one distinct stage is present, consider each stage as one
spawning and handle separately as described below. If spawning occurred that
morning, embryos will typically be undergoing late cleavage, and determination of
fertility (number of embryos/number of eggs x 100) is easily achieved. Occasionally,
spawning occurs in the afternoon and embryos will be close to tailbud stage by the
following morning. A detailed embryonic staging sequence for fathead minnows is
provided in U.S. EPA (1996b). Infertile eggs are opaque or clear with a white dot
where the yolk has precipitated; viable embryos will be clear for the first 36 to 48 h
until they reach the eyed stage. An alternative to the microscopic approach to
determining fertility is to enumerate eyed embryos at this time. If the latter technique
is used, the embryos need to be held in a system apart from the adults in a water bath
with aeration (U.S. EPA 1987) or placed in incubation cups (U.S. EPA 1982). The
fertility rate in control animals generally will exceed 95%.
e. Embryo Hatch (Optional Endpoint). Regardless of the method used to determine
fertility, if information concerning hatching success and/or subsequent development
is desired, the embryos need to be maintained in incubation chambers. Depending on
study objectives, embryos can be held in clean water or water containing the same
concentration of chemical to which the parents were exposed to continue an Ft
, . 53 ',
-------
generation exposure. Incubation cups can be made from glass cylinders or jars with
the bottoms cut off and nylon or stainless steel screen glued to the bottom with
silicone adhesive (U.S. EPA 1982). The cups can be suspended in a tank with an
oscillating water level to ensure that the embryos are always covered and that water
regularly flows into and out of the cups without creating excessive turbulence.
Alternatively, they can be held in incubation cups containing clean water under static
conditions. If this is done, the incubation cups should be held in a constant-
temperature water bath and water inside the cup renewed daily. Discrete spawns
should be screened with a dissecting scope to remove empty or opaque shells, or
abnormal embryos, and a subsample (e.g., 50) should be selected at random and
placed in the incubation cup. At 25°C, untreated animals will hatch in approximately
4 to 5 d. Embryos should be inspected daily and any dead animals counted and
removed. A small container should be
placed under the incubation cup prior to
removal from the water bath for exami nation under a dissecting scope. Potential
endpoints include time to complete hatzh, total number of embryos hatched, and
number of normal larvae at hatch. The; hatching rate of control animals typically is in
i
I
the range of 95 to 98%. ;
f. Larval Survival and Morphology. (Optional Endpoint). Newly hatched fry should be
observed for general physical appearance and behavior. Healthy fry should actively
swim about the incubation cup. Gross morphological abnormalities, described by
Sharp (1991), which may be observed include lordosis, scoliosis, kyphosis, retarded
54
-------
swim bladder development, edema, and craniofacial abnormalities. Survival of the
larvae can easily be assessed through yolk sac absorption (-1 to 2 d post-hatch at
25°C); if estimates of survival are required after this, the animals must be fed
(generally live Anemia nauplii; U.S. EPA 1987). An extended observation period
may be necessary to detect chemicals that cause delayed mortality, such as Ah
receptor agonists (Elonen et a/., 1908). Fry are fed live brine shrimp nauplii (< 24 h
old) until approximately 30 d of age (U.S. EPA 1987), possibly supplemented from
20 d of age with powdered dry food (Lange et al 2001).
g. Secondary Sex Characteristics. Observations of physical appearance of the adults
should be made over the course of the test and at conclusion of the study. From the
perspective of screening EDCs, characteristics of particular importance include body
color (i.e., light or dark), coloration patterns (i.e., presence or absence of vertical
bands), body shape (i.e., shape of head and pectoral region, distension of abdomen),
and specialized secondary sex characteristics (i.e., number and size of nuptial
tubercles, size of the dorsal pad and ovipositor). Notably, chemicals with certain
MOA can cause the abnormal occurrence of certain secondary sex characteristics in
the opposite sex; for example, androgen receptor agonists, such as methyltestosterone,
can cause female fathead minnows to develop nuptial tubercles (Smith 1974; Ankley
et al. 2001), while some estrogen receptor agonists may decrease number or size of
nuptial tubercles in males (Miles-Richardson et al. 1999a; Harries et al., 2000; Lange
et al. 2001). Any abnormalities in appearance should be documented and, to the
55
-------
extent possible, quantified (e.g., countijng tubercles). It may be advantageous to
r
photograph or videotape the appearance of the fish near the end of the test. Because
some aspects of appearance (primarily Color) can change quickly with handling, it is
important that qualitative observations be made prior to removal of animals from the
i
test system. This type of qualitative assessment might be enhanced through the use of
t
photographs or videotape. Other endpoints, such as the number and size of nuptial
i
tubercles, can be quantified directly or in preserved specimens.
i
An example of a data form to record all pertinent sample data at the conclusion of the
test is provided (Appendix A, Form A-J3).
i
!
h. Blood Collection. At conclusion of the' 21-d test, the fish should be netted from the
exposure chambers one at a time and anaesthetized with MS-222 (100 mg/L, buffered
with 200 mg NaHCOj/L). After anaesttietization, the caudal peduncle should be
partially severed with a scalpel blade and blood collected from the caudal vein/artery
with a heparinized microhematocrit caj illary tube (Figure 6). After the blood has
been collected, the plasma is quickly isolated by centrifugation for 3 min at 15,000 g.
If desired, percent hematocrit can be determined following centrifugation. The
plasma portion is then removed from the microhematocrit tube and stored in a
centrifuge tube with 0.13 units of aprotinin (a protease inhibitor) at -80°C until
determination of vitellogenin and sex steroid concentrations can be made. Depending
56
-------
on size of the fathead minnow (which is sex-dependent), collectable plasma volumes
generally range from 20 to 60 ul per fish (Jensen et al. 2001).
Figure 6. Fathead minnow blood collection process.
! (
i. Vitellogenin. Vitellogenin is a phospholipoglycoprotein precursor to egg yolk protein
that is synthesized in the liver of sexually-mature females of all oviparous species; the
production of Vitellogenin is controlled by interaction of estrogens, predominantly
p-estradiol, with the estrogen receptor (Sumpter and Jobling 1995; Kime 1998).
Significantly, males maintain the capacity to produce Vitellogenin in response to
stimulation with estrogen receptor agonists; as such, induction of Vitellogenin in
males has been successfully exploited as a biomarker specific for estrogenic
compounds in a variety of fish species, including the fathead minnow (Kramer et al.
57
-------
1998; Panter et al 1998; 2002; Parks et al 1999; Tyler et al 1999; Harries et al
2000; Korte et al 2000; Ankley et al 2001; LSnge et al 2001).
Different methods are available to assess vitellogenin production in fish; a
measurement technique that is both relatively sensitive and specific is determination
of protein concentrations in plasma via enzyme-linked immunosorbant assay
(ELTSA). In performing quantitative EIISAs with fathead minnow samples, some
have utilized polyclonal or monoclonal vitellogenin antibodies prepared using protein
i
from other fish species that cross-reacts with fathead minnow vitellogenin, (Tyler et
al 1999; Harries et al 2000). Others have used fathead minnow-specific polyclonal
vitellogenin antibody and purified fathead minnow vitellogenin protein for the EOS A
(Parks et al 1999; Korte et al 2000). [Using the fathead minnow-specific approach,
detailed instructions for the measurement of fathead minnow vitellogenin are given in
i
Appendix B. j
i
i
I
j. Sex steroids. Plasma concentrations of p-estradiol, testosterone, and
11-ketotestosterone can be determined; using radioimmunoassay (RIA) techniques
i
optimized for the relatively small sample volumes obtained from the fathead minnow
t
(Jensen et al 2001). There is emerging literature documenting concentrations of
i
these sex steroids in the fathead minnow under both normal conditions and after
treatment with various EDCs (Giesy et al 2000; Makynen et al. 2000; Ankley et al
2001; 2002; Jensen et al 2001; 2002).
In mature females, both p-estradiol and
58
-------
testosterone are easily measured, While 11-ketotestosterone concentrations typically
are low to non-detectable. In mature males, concentrations of 11-ketotestosterone are
two- to five-fold greater than those of testosterone, and p-estradiol concentrations are
low to non-detectable (Jensen et at. 2001).
Detailed instructions for the measurement of fathead minnow plasma sex steroids
using RIA are given in Appendix C.
k. Gonadosomatic Index. An assessment of reproductive status can be determined by
measurement of the gonadosomatic index (GSI). The GSI is the weight of the ovaries
or testes relative to the total body weight of the fish (GSI = 100 x gonad weight
(g)/body weight (g); Grim and Glebe 1990). After removing blood, wet weight
should be measured on individual fish to the nearest 0.01 g. A ventral incision
through the body wall from the anus to the isthmus should be made and both ovaries
in females and both testes in males carefully removed with fine forceps and placed
into a pre-weighed weighing pan. Gonad wet weight should be measured to the
nearest 0.1 mg.
Typical GSI values for the fathead minnow range from 1 to 2% for males and 8 to
13% for females (Smith 1978; Jensen et al. 2001). In fractional spawning fish, like
the fathead minnow, the ovaries undergo rapid cyclical changes over relatively short
periods of time as successive batches of eggs are produced. Female GSI varies
59
-------
significantly as a function of spawning interval, with the smallest values occurring
just after spawning followed by an inc pease of 45% within 3 to 4 d, just prior to the
next spawning event (Jensen et al. 20C
1). This means that the ovaries in breeding
females can vary considerably between individuals within treatments at any one point
i
in time. This could make identifying alterations in GSI as a result of exposure to
EDCs difficult if the normal variation associated with the reproductive cycle exceeds
effects associated with exposure to thej test chemical. Nevertheless, significant effects
I
on GSI have been reported in fathead minnows following exposure to various EDCs
mil
(Panter et al. 1998; Makynen et al. 2000; Ankley et al. 2001; 2002).
1. Gonadal histology. Histological evaluation of the gonads is an important component
of the test because it may reveal specific alterations in the gonads leading to possible
\
impacts on reproductive output and also provide possible insight into the mechanism
of action of potential EDCs. Several r scent studies involving exposure of fathead
minnows to EDCs have included an examination of gonadal histology. Most, but not
all, of these studies have considered ef Sects on both ovaries and testes. Ovaries of
females exposed to strong estrogen receptor agonists such as p-estradiol or
ethinylestradiol exhibit fewer mature, and more atretic, follicles (Miles-Richardson et
al 1999a; Lange et al 2001). Kramer et al. (1998) and Miles-Richardson et al
(1999a) noted that P-estradiol exposure resulted in a sustained increase in plasma
vitellogenin in females. They suggested that sustained abnormally high vitellogeinin
levels interfere with final maturation and release of oocytes from the ovary, possibly
60
-------
by inhibiting gonadotropin release by the pituitary. Testicular effects varied,
depending on concentration, from slight degeneration, especially involving
spermatozoa, to clear atrophy and appearance of ova-testes. Moreover, Miles-
Richardson et al. (1999a) reported that a concentration-dependent proliferation of
Sertoli cells containing remains of spermatozoa accompanied degeneration of
spermatozoa. Exposure to methoxychlor, a weak estrogen receptor agonist, resulted
in increased follicular atresia (Ankley et al 2001), a finding consistent with a
, I • i . ',
substantial reduction in fecundity during a 21-d exposure. Exposure to nonylphenol,
another weak estrogen receptor agonist, resulted in significant necrosis of germ cells
and spermatozoa accompanied by hyperplasia and hypertrophy of Sertoli cells
containing germ cell remnants (Miles-Richardson et al. 1999b). Examination of the
ovaries of females exposed to the strong androgen receptor agonist,
methyltestosterone, revealed no postovulatory follicles (corpora lutea) (Ankley et al.
2001). Instead, maturation of younger follicles was suppressed and there were
numerous pre-ovulatory atretic follicles, consistent with the observation that
spawning ceased immediately following exposure to the androgen.
Methyltestosterone-exposed testes differed from controls in that they appeared to be
stimulated to near exhaustion of germinal epithelial stages. The germinal epithelium
was much thinner, and spermatogenic activity more scattered than in control testes
(Ankley et al. 2001). The effects of two putative anti-androgens also have been
evaluated. Makynen et al. (2000) observed significantly reduced oocyte diameters
following exposure to vinclozolin. That is, the ovaries exhibited retarded maturation,
61
-------
especially regarding yolk deposition in oocytes. Similarly, exposure to flutamide
resulted in more early-stage follicles and an increase in atretic follicles (Jensen et al.
2002). Both of these factors could indicate decreased egg production via a delay in
oocyte maturation resulting in fewer eggs being produced in a given time or a greater
than average number of follicles undergoing resorption before reaching maturity.
Exposure to the aromatase inhibitor, fadrozole, resulted in a near (lowest
concentration) or complete (higher concentrations) absence of corpora lutea,
I
increased numbers of pre-ovulatory atiretic follicles, and regression to an earlier
developmental stage (higher concentrations) (Ankley et al. 2002). Judging by
i
histological criteria, fadrozole appeared to reduce fecundity by preventing maturation
of oocytes.
Routine histological procedures can be used to assess fathead minnow gonads.
Appendix D discusses two approaches to fixation and embedding appropriate for
fathead minnow gonadal histology: a traditional paraffin-based approach and a more
i
modern methacrylate-based histological procedure. Either technique is acceptable in
the context of the test described in the
document. An assessment of normal gonadal
histology and routine methods for evaluating EDC-induced histological changes in
reproductively-mature fathead minnow gonads is given in Appendix E.
i
!
i
m. Tissue Residues (Optional Endpoint). Adult tissue residue analysis verifies that the
!
chemical entered the fish, and also offers an opportunity to calculate chemical-
62
-------
specific bioconeentration or bioadcumulatioii factors. Further, the relationship of
specific tissue residues to toxic effects can be valuable in terms of across-chemical,
across-species, and laboratory-to-field extrapolations (Jarvinen and Ankley 1999).
Because of potential among-fish variations in tissue residues, individual fish should
be analyzed whenever possible. If it is necessary to pool samples to obtain adequate
tissue mass for analyses, it is recommended that males and females from any given
treatment be kept separate because of possible gender-related differences in
bioconeentration (Makynen et al. 2000; Ankley et al. 2001). For some types of test
chemicals (e.g., nonionic organics), measurement of organism lipid content could
prove useful for assessing bioconcentration/bioaccumulation data (Ankley et al.
2001).
n. Other Endpoints (Optional). Cytochrome P450 aromatase (CYP19) catalyzes a key
step in the conversion of C19 androgens to CIS estrogens. The ovarian form of this
enzyme is the major source of circulating 0-estradiol in females (Mommsen and
Walsh 1988). In fish, both males and females also have significant aromatase activity
in brain tissue (Gelinas et al 1998; lyfelo et al. 1999). Inhibition of this critical
enzyme could result in reduced levels of p-estradiol and alter the expression of many
genes controlled by the estrogen receptor (e.g., vitellogenin production).
Activity of the enzyme can be measured using a simple assay in which radiolabeled
androstenedione releases tritiated water during aromatization and conversion to
' 63 '• '
-------
estrone (Thompson and Siiteri 1974). [ Several methods have been reported in the
j.
literature for measuring this activity in fish tissue (Pelissero et al 1996; Chang et al.
1999; Melo et al 1999; Shilling et al 1999; Kitano et al 2000). Because of the small
size of the fathead minnow, a modification of the microassay of Melo et al (1999)
has been used with brain tissue from fathead minnows (Ankley et al 2002). In the
latter study, adults exposed to the aromatase inhibitor, fadrozole, exhibited a
i
concentration-dependent decrease in brain aromatase activity.
The expression of the aromatase gene has also been determined by measuring gonadal
I
aromatase mRNA (Kitano et al 2000; Scholz and Gutzeit 2000). Determination of
i
ovarian aromatase mRNA levels is als'o possible in individual fathead minnows.
Using the polymerase chain reaction (PCR) and consensus primers designed from the
DNA sequence of aromatase genes in other fish, an amplified segment of the fathead
minnow aromatase cDNA was obtained (GenBank accession number AF288755).
This segment of DNA was used to generate a biotin-labeled antisense RNA probe for
use in measuring mRNA levels via a nbonuclease protection assay (RPA) (J. Korte,
unpublished data). The same sequence information could be used in a quantitative
real-time PCR method. Measurement
particularly useful in fathead minnow
of aromatase expression via mRNA levels is
ovaries since the small size of the tissue makes
direct enzyme activity measurement challenging in individual fish. It is not currently
known if changes brought about by aromatase inhibitors, such as fadrozole, will alter
the level of expression of the gene at t le mRNA level.
\
64
-------
Since the effects of androgens must be manifested through binding to the androgen
receptor, some means of measuring changes in the level of the receptor could be
effective in identifying EDCs. Plasma p-estradiol levels are known to be capable of
up-regulating the level of the estrogen receptor in fish liver (Pakdel et al. 1991;
Andreassen and Korsgaard 2000). Although regulation of the androgen receptor is
not well understood, there is some evidence of tissue-specific up- and down-
regulation by 11-ketotestosterone (Todo et al. 1999). pbservations with female
fathead minnows have shown the appearance of nuptial tubercles when exposed to the
androgen receptor agonist, methyltestosterone (Ankley et al. 2001). These tubercles
are normally only found in sexually-mature males. Since the appearance of tubercles
would seem to be a response mediated through the androgen receptor, it is possible
that the appearance of male secondary sex characteristics in female fish involves
changes in the level of the androgen receptor. Using PCR and consensus primers
designed from the DNA sequence of androgen receptor genes in other fish, an
amplified segment of fathead minnow androgen receptor cDNA was obtained
(GenBank accession number AJ277866). This information was used to prepare
biotin-labeled antisense RNA probes for use in a RPA to measure androgen receptor
mRNA (J. Korte, unpublished data). Conversion to a quantitative real-time PCR
method could also be easily accomplished.
65
-------
I. Treatment of Results
1. Overview
The reproduction test generates quantitative data on adult survival, GSI, plasma
concentrations of vitellogenin and sex steroids, fecundity, fertility, and, if desired, embryo
hatch and larval survival. The null hypothesis that no differences exist between the control
fish and those exposed at the different treatment levels should be statistically evaluated for
each test endpoint. Both types of statistical error (i.e., Types I and n) should be clearly
specified in the presentation of the statistical results.
i
f
Time-dependent data (i.e., fecundity and fertility) potentially require the most complex
analysis because this information is collected
daily both before and after exposure to the test
chemical. These data can be analyzed in many ways. Some methods are more appropriate
than others for this type of data, which can be highly variable especially when the mean
fecundity is large, temporally correlated, or skewed. Mean fecundity per period (i.e.,
pre-exposure or post-exposure) and test chamber can be relatively variable, and variance
usually increases with mean fecundity. Also,
the correlation between pre-exposure and
post-exposure mean fecundity from the same test chamber may be low. The recommended
analysis is based on a summary statistic for ezch test chamber, either the mean post-exposure
fecundity or the ratio of the post-exposure to pre-exposure fecundity.
66
-------
Data on other endpoints do not present the complexities of the time-dependent data (i.e.,
fecundity, fertility); they are measured only once at the termination of the experiment (adult
survival, GSI, and plasma vitellogenin and sex steroid concentrations).
The use of the suggested statistical methods for routine data analysis does not require the
assistance of a statistician. However, interpretation of the results may become problematic
because of the inherent variability and occasionally unavoidable anomalies in the data. If the
data appear unusual in any way, or fail to meet the necessary assumptions, a statistician
should be consulted. Analysts who are not proficient in statistics are strongly advised to seek
the assistance of a statistician before selecting the method of analysis and interpreting the
results.
2. Analysis of Fecundity Data1
a. Data preparation and adjustment for mortality. The experimental period is divided
into three time phases. The first seven days should be considered an acclimation
period; omitting these data from subsequent analyses may reduce variability. The
remaining period until the treatment is imposed is considered the pre-exposure
period. Observations from day 1 of exposure until the end of the experiment are the
post-exposure period.
Although this section discusses fecundity data specifically, most of the
considerations and approaches are also germane to fertility and, if collected, hatch
data.
67
-------
The mean (± SD) daily fecundity during the pre-exposure and post-exposure periods
should be calculated for each test chamber. The daily observations of fecundity can
be highly variable; calculating the mean fecundity reduces some of this temporal
variability. Also, the experimental unit is the test chamber and analysis of the data is
!
much simpler when summarized into one value per chamber. That single value per
chamber could be the post-exposure fecundity, the ratio of fecundities, or some other
\
function of the fecundity. The choice [should be specified before the analysis is
j
j
conducted.
When there is no mortality, the mean daily fecundity is an appropriate measure of
spawning success. These experiments are designed to estimate sub-lethal effects so
i
the default assumption is that test concentrations are low enough so that there should
be no treatment-related mortality. However, random deaths may occur, and loss of
one or more females may decrease the; mean daily fecundity in that chamber. Two
i
i
approaches could be used to adjust for! random mortality.
1) Compute the mean fecundity on a per live-female per day basis. Compute the
I
total number of eggs produced during the pre-and post-exposure periods.
Then, compute the total number of female reproductive days during each
period. The corrected mean is the ratio of the total number of eggs to the
i
number of female reproductive! days. (
i
i
68
-------
2) Omit fecundity data from any test chamber with mortality.
The first approach uses data from all chambers, but assumes that the mortality of a
female has no effect on any other fish, that the mortality was not caused by exposure
to the test chemical, and that there was no period of reduced fecundity prior to death.
The second approach only assumes that the mortality was not caused by the test
chemical, but it results in a smaller sample size and smaller statistical power. All the
available biological information should be used to choose the more appropriate
adjustment. In instances where the death of a male fish occurs, the second approach
may be more appropriate.
The fecundity data should be plotted, both as a preliminary step to help detect
anomalies and unsuspected trends or patterns in the responses, and as an aid in the
interpretation of the results. These plots can also be used to visually check the
validity of some assumptions that are necessary for several statistical analyses that
may be used.
b. Choice of analysis. The summary statistics* one for each test chamber in the
experiment, can be analyzed by many different statistical procedures. One approach
is to use general characteristics of the data to identify an approach and possible
11 , '
alternatives. All tests are conducted on the same organisms, in a similar manner, and
! ;
measure the same response, mean fecundity. The mean fecundity will not be the
69
-------
same in each experiment, because different chemicals have different effects.
However, every one of the tests provide information about the characteristics of the
data that influence the choice of analysis.
Consideration of possible approaches
for analyses can be organized into four
decisions; recommendations for each [decision should be based on general
characteristics of the data. The recommendations below are tentative. 'They may
change as this test is conducted more frequently, and general characteristics of the
resultant data become better known.
1) Choice of response: It is possible
to use mean fecundity during the post-exposure
period or use the ratio of the post-exposure and pre-exposure fecundity. A third
option is to use the difference between post-exposure and pre-exposure mean
fecundity. The choice of response affects the statistical power of the analysis. If
the correlation between pre-exposure and post-exposure fecundity is high or
moderately high, analysis of the n tio is more powerful. If the correlation is low,
analysis of the post-exposure data
alone is more powerful. Often, because of the
large inherent variability, the observed correlation will be relatively small, so it
may be most common to analyze the post-exposure mean.
2) Choice of hypothesis to be tested: It is possible to test the hypothesis of equal
i
means, or test for a monotonic concentration (or dose) response, or test whether
70
-------
specific treatments differ from the control. The traditional hypothesis, tested by
analysis of variance (ANOVA), is equal means. This is appropriate for any set of
treatments, but it ignores extra information available when the treatments are a
sequence of test concentrations. If the concentration (dose)-response curve is
expected to be monotonic, a more powerful test is that of ordered alternatives.
This tests the hypothesis that fic ;> IIL ;> fiH, where ft, is the mean fecundity in the
ordered treatments, ranging from control to higher concentrations. If desired, this
test can be followed by comparisons of specific concentrations to the control
using an adjustment for multiple comparisons.
3) Choice of parametric (ANOVA) or nonparametric analysis: This choice affects
the statistical power and validity.of the a level. It is also influenced by the ease of
computation for a test. If the data satisfy the ANOVA assumptions (data are
independent, normally distributed, and have equal variances), ANOVA is slightly
more powerful than nonparametric alternatives. When the experimental design is
balanced, i.e. equal number of replicates of all treatments, violations of the
normality and equal variance assumptions have little effect on the a level, but
decrease the statistical power. Nonparametric ordered alternatives tests require
moderately complicated calculations and are not available in commonly-used
computer software. A good nonparametric test of ordered alternatives, the
Jonckherre-Terpstra test, is available in SAS (SAS Institute, Gary, NC) or SPSS
(SPSS, Inc., Chicago, IL); if hand calculations are necessary, they are straight-
71
-------
forward. Parametric multiple comparisons tests (e.g. Dunnett's) are widely
available in computer software; nonparametric multiple comparisons are not, but
again hand computations are strai ^it-forward. The recommended nonparametric
test is the Jonckherre-Terpstra tes , followed by multiple comparisons using the
nonparametric analog of Scheffe's
method.
4) Choice of transformation: ANOyA is a parametric procedure based on the
assumptions that observations within treatments are independent and normally
|
distributed, and the variance of the data are homogeneous. These assumptions
must be checked prior to using this approach to determine if they have been met
(tests for validating these assumptions are discussed in I.3.c). When the
i
assumptions of normality and/or homogeneity of variance are not met,
transformations of the data may remedy the problem, so that the data can be
analyzed by parametric, rather than nonparametric procedures. Often, but not
always, the same transformation will do both. When the data include replicates of
the same treatment, a Box-Cox pi
)t can be used to choose between specific
transformations to equalize variances (Box et al. 1978). The mean and
within-group standard deviation are calculated for each group of observations.
Then, y=log(SD) is plotted against x=log(mean). The slope of the linear
i
regression indicates the transformation that best controls the variance-mean
relationship. A slope near 0 indicates no transformation is required (no
variance-mean relationship). A slope near 0.5 indicates a square root
72
-------
transformation is needed, while a slope near 1 indicates that a log transformation
is most appropriate. Often, the estimated slope will be near 1, suggesting the use
of a log(x) transformation. On occasion, this may need to be modified slightly
because fecundity may be zero and log(0) is undefined. The recommended
transformation for the mean post-exposure fecundity is log(x+l) where x is the
mean post-exposure fecundity. The recommended transformation for the ratio is
log((x+l)/(y+l)), where x is the mean post-exposure fecundity and y is the mean
pre-exposure fecundity.
The choice of transformation is not an issue if a nonparametric approach is
chosen. The suggested nonparametric tests are based on the ranks of the
observations. All the transformations considered here are monotonic; the ranks of
the transformed variables are identical to the ranks of the original variables.
Hence, the nonparametric analysis of transformed variables is identical to the
nonparametric analysis of the untransformed variables.
c. Analysis of Variance. If a parametric analysis is utilized, one-way ANOVA can be
used to test the hypothesis of no differences between treatment means. If the
response is chosen to be the mean post-exposure fecundity, it should be calculated for
each chamber, possibly adjusting for mortality if any occurred. The post-exposure
mean fecundity in test chamber i, x,, should be transformed to log £+1). If the
73
-------
response is chosen to be the ratio, the transformed ratio should be calculated for each
r
[
[
chamber as log (=77), where yt is the mean pre-exposure fecundity in chamber i.
For either response, treatment means and the F-test of equal means are calculated
using standard computer software for one-way ANOVA. Following the ANOVA,
comparisons between specific treatments can be made by fitting a concentration
(dose)-response curve (e.g., a linear cpntrast), by testing each concentration against
the control (using Dunnett's Multiple Comparisons Procedure), or comparing all pairs
of treatments (using Tukey's Multiple! Comparison Procedure). ANOVA is discussed
in further detail in Section 1.3.
d. Jonckherre-Terpstra test of ordered alternatives. The Jonckherre-Terpstra test is a
nonparametric test of the hypothesis t lat the mean fecundities in each treatment
follow a specified order. When there
and "High" concentrations), and three
are three treatments (e.g., "Control," "Low,'
treatment means (/ic, /iL, and f%), the
hypothesis being tested is that nc z fa z /%, with at least one strict inequality, against
the null hypothesis that HC — ^L — /%•
When the treatments are a sequence of
increasing concentrations, with the concentration-response curve expected to be
I '
monotonic, a test of ordered alternatives is more powerful than the usual test of equal
means (e.g., a one-way ANOVA or the nonparametric Kruskal-Wallis test).
74
-------
The Jonckheire-Terpstra test can be calculated by S AS PROC FREQ (S AS Institute,
Gary, NC), which can also calculate the exact p-value, the SPSS EXACT module
(SPSS, Inc., Chicago, IL), and perhaps other statistical packages. It is not a test that is
widely available in all statistical software. If a statistical package can calculate the
Mann-Whitney (or equivalently the Wilcoxon's rank sum test) for a pair of
treatments, the Jonckherre-Terpstra test statistic can be obtained from the results of
Mann-Whitney tests on all pairs of treatments. If no nonparametric test is available,
the Jonckherre-Terpstra test statistic can be calculated by hand.
Hand calculation of the Jonckherre-Terpstra is described and illustrated in some
nonparametric statistics books (e.g., Daniel 1978). The statistic is calculated by
considering all unique pairs of treatments one pair at a time. If there are three
treatments, there are three pairs of treatments. Label each pair so that the higher
concentration (with the lower expected fecundity) is treatment X and the lower
concentration (with the higher expected fecundity) is treatment Y. Consider all
possible ways that an observation in treatment X can be paired with an observation in
treatment Y. If there are four observations in each treatment, there are sixteen
possible pairings of observations. Count the number of pairs where the observation
from treatment X is smaller than the observation from treatment Y. If two values are
tied (i.e., the observations in treatment X and Y have the same fecundity), add one
half for each tied pair to the count. Add these counts up over all pairs of treatments.
75
-------
This sum is J, the test statistic for the Jonckherre-Terpstra test. A large value
indicates that the fecundity tends to be larger in treatments with lower concentrations.
Exact critical values of the Jonckherre-Terpstra test statistic for a variety of sample
sizes are computed in Odeh (1971) ani tabulated in Daniel (1978). Critical values for
the sample sizes and number of treatments used or potentially used in the test are
given in Table 3. The test rejects the hull hypothesis if the observed test statistic is
i
larger than or equal to the tabulated value. The sample sizes for these tests are
sufficiently small that the asymptotic Chi-square distribution should not be used.
76
-------
Value of J
33
34
35
36
37
38
39
40
41
42
43
44
45
62
63
66
67
69
70
72
73
75
76
n=4, fc=3
0.110
0.084
0.063
0.046
0.033
0.023
0.015
0.0099
0.0062
0.0037
'
.n=3,k=4
0.117
0.091
0.069
i 0.052
0.037
0.027
0.018
0.012
0.0080
0.0050
r
( ,
n=4,*=4
0.1058
0.0895
0.0514
0.0420
0.0272
0.0215
0.0123
0.0100
0.0056
0.0041
Table 3. P-values corresponding to specified values of /, the test statistic for the
Jonckherre-Terpstra test of ordered alternatives (Daniel 1978). Values are provided
for n=4 replicates of fc=3 treatments, n=3 replicates of fc=4 treatments, and n=4
replicates of k=4 treatments. Values of / for n=4, k=4 are included only when the
p-value brackets a commonly used a level (e.g., the two values J=66 and J=67 bracket
the 0.05 p-value).
77
-------
e. Kruskal-Wallis test of equal treatment medians. An alternative to the
i
Jonckherre-Terpstra approach is the Kruskal-Wallis test, the nonparametric
equivalent of the one-way ANOVA. This test is generally available in statistical
software packages. Because the sample sizes are small, exact p-values should be used
instead of the asymptotic Chi-square distribution. Tables of exact p-values are
available in many nonparametric statistics texts and some collections of statistical
tables. Exact p-values can also be computed using S AS or SPSS (Exact module).
Critical values for two common choices of a level are given in Table 4.
f. Nonparametric multiple comparisons between treatments. Nonparametric multiple
i
comparisons between means are based on the mean ranks from the Kruskal-Wallis
test. These values are usually reported in the output from Kruskal-Wallis analyses.
The recommended procedure uses Scheffe's approach, adapted for the analysis of
ranked data. This is appropriate for any set of comparisons between groups,
including comparisons of each non-zero group to the control. For each comparison,
calculate the absolute value of the difl erence in mean ranks, T= \-Ri-Rj\. The
difference between these two groups is significant if T exceeds the critical value given
by:
12
which reduces to
78
-------
V-r
when the two groups have the same sample size. The term h^_ l is the critical value
for the a level Kruskal-Wallis test with k treatments where k is the number of groups
and n{ is the number of observations in group i. When the sample sizes are equal, n is
the number of observations per group. N is the total number of observations in the
entire experiment. A subset of the critical values for multiple comparisons is given in
Table 4.
Kruskal-Wallis test
n
4
3
4
k
3
4
4
a = 5%
5.692
7.000
7.235
a=l%
7.654
8.538
9:287
Multiple comparison
a = 5%
6.082
7.791
9.055
a = l%
7.052
8.603
10.26
Table 4. Critical values for the Kruskal-Wallis test (Daniel 1978) and Scheffe's
nonparametric multiple comparisons approach with experiment-wise error rates of
5% and 1%. Values are provided for w=4 replicates of k=3 treatments, n=3 replicates
of k=4 treatments, and n=4 replicates of fc=4 treatments.
79
-------
3. Analysis of Other Endpoints
a. Data preparation. Data on other endpbints do not provide the complexities of the
t
i
fecundity or fertility data; they are measured only once at the end of the experiment
(adult survival, GSI, and plasma vitellogenin and sex steroid concentrations). The
I
data should be plotted, both as a preliminary step to help detect anomalies and
unsuspected trends or patterns in the responses, and as an aid in the interpretation of
the results. These plots can also be used to visually check the validity of some
assumptions associated with several statistical analyses that may be used. Scatter
plots of two or more variables may demonstrate the relationships among the variables,
so that correlations can be observed and interactions studied.
The experimental unit in these tests is
reduced to chamber means before use
the test chamber; however, most of the
endpoints are measured using individual fish. Fish-specific observations must be
in a statistical test of treatment differences.
b. Outliers. Outliers are inconsistent or questionable data points that appear either high
or low compared to the general trend exhibited by the majority of the data. Outliers
can be detected by tabulating the data, plotting, or an analysis of the residuals.
i
Explanations should be sought for any questionable data points. If the outlier cannot
be linked to an explanation such as faulty equipment or human error, the outlier could
i
be the result of a real effect. Without an explanation, data points should be discarded
i
i
80
-------
only with extreme caution. If there is not an explanation, the analysis should be
performed both with and without the outlier, and the results of both analyses reported.
Outliers increase the variability within a sample and, therefore, decrease sensitivity of
the statistical tests used.
c. Choice of analysis. The process of selecting appropriate statistical tests should follow
a decision tree similar to the one illustrated in Figure 7. This approach uses tests of
normality and tests of equal variance, and experimental characteristics to guide the
selection of the appropriate analysis, the choice of test is based on the characteristics
of each specific data set. Dunnett's Procedure and the Mest with Bonferroni's
adjustment are parametric procedures based on the assumptions that the observations
within treatments are independent and normally distributed, and that the variance of
the observations is homogeneous across treatments. These assumptions should be
checked prior to using the tests to determine if they have been met. If the tests fail, a
nonparametric procedure such as Steel's Many-one Rank Test or Wilcoxon' s Rank
Sum Test may be more appropriate.
!
After examining the plots and descriptive statistics, assumptions of normality and
homogeneity of variances among groups are tested. Normality can be tested using
Shapiro-Wilks Test, among others. In general, if the data fail the test for normality, a
transformation such as to log values may normalize the data. Homogeneity of
variances across groups can be tested using Bartlett's Test, among others. In using
81
-------
this test, it is assumed that the data are normally distributed. If data display or tests
demonstrate that variance is not homogeneous across treatments, then variance-
stabilizing transformations of the data may be necessary. The arcsin, square root, and
•i ' ••
logarithmic transformations are often jused on dichotomous, count, and continuous
I
data, respectively. |
The choice of a particular analysis and the ability to detect departures from the
assumptions of the analysis, such as the normality of the data and homogeneity of
variance, is dependent on the number of replicates. One concern is that if the sample
sizes are too small, the tests of equal variances and normality may have low power to
detect unequal variances or non-normality, unless the deviations are very large.
i
Because the sample sizes are relatively small under the current recommended
experimental design, the data from a single study may provide relatively little
information to test normality or equal
variances. The statistical consequence is that
tests of normality and unequal variances may have low statistical power.
If the assumptions are met, the data can be subjected to ANOVA followed by
Dunnett's Multiple Comparison Procedure for comparing each of the treatment means
with the control mean to determine if any of the concentrations differ from the
control. It is based on the assumptions that the observations are independent and
t
i
normally distributed and that the variance of the observations is homogeneous across
all treatments. Dunnett's Procedure can only be used when the same number of
82
-------
replicates (test chambers) have been used at each treatment. In cases where the
number of replicates for different treatment levels are not equal, a f-test with
' ' - \
Bonferroni's adjustment can be used as an alternative to Dunnett's Procedure. The
Mest with Bonferroni's adjustment is based on the same assumptions of normality
and homogeneity of variance as Dunnett's Procedure.
If, after suitable transformations have been carried out, the assumptions of ANOVA
have not been met, nonparametric techniques should be used. Steel's Many-one Rank
Test is a nonparametric test for comparing treatments with a control and is an
alternative to Dunnett's Procedure. If the number of replicates (test chambers) are not
equal, Wilcoxon's Rank Sum Test with Bonferroni's adjustment should be used.
If both solvent and dilution water (non-solvent) controls are included in the test, they
should be compared using a Mest for count and continuous data and Fisher's Exact
Test for categorical data. If the difference between controls is not statistically
significant, then all control data can be combined for the remaining analyses.
Various software packages are available to perform these analyses such as SAS (SAS
Institute, Gary, NC), SigmaStat or SYSTAT® (SPSS, Inc., Chicago, IL), ToxCalc
(TidePool Scientific Software, McKinleyville, CA), or TOXSTAT® (Western
Ecosystems Technology, Inc., Cheyenne, WY).
83
-------
Adult Survival
Fertility
Embryo Hatch
Larval Survival
GSI
Vitellogenin
Plasma Steroids
Fecundity
Hypothesis Testing
Using ANOVA
Data
Transformations
Shapiro-Wilfc's Test
Normal Distribution
Homogeneous
Variance
Non-Normal Distribution
Bartlett's Test
Equal Number
of Replicates
No
Yes
T-Testwith
Bonferroni
Adjustment
Dunnett's Test
Heterogeneous
Variance
Equal Number
of Replicates
Yes
No
Wilcoxon's Rank
Sum Test with
Bonferroni
Adjustment
Figure 7. Flow chart for statistical analysis of data from the 21-d reproduction
test with the fathead minnow.
84
-------
4. Summarization
Following completion of the statistical analyses and interpretation of results, the mean
and standard deviation for each endpoiht should be incorporated into a summary table or
figure. Statistical significance, as well as the statistical tests used, level of significance,
and sample size should be indicated.
J. Test Report
A test report form must be completed for each chemical evaluated. An example of a report form
*• t
for the 21-d reproduction test is provided in Appendix A (Form A-4).
1. Test Chemical
The report must include a detailed description of the test substance, including information on
its CAS number, source, lot number, and purity. Additional information should be provided,
when available, such as its solubility in water, K,,w, vapor pressure, etc. Toxicity to the
fathead minnow (or other fish species) should also be reported along with pertinent
information from the range-finding test.
85
-------
2. Test Animals
I
i
t
Information must be provided on the fathead minnows used in the test. This information
i ^
must include the source of the fish, age, and condition of the fish at the initiation of the test,
\
the pre-exposure reproductive performance, the acceptability of the most recently performed
reference toxicity test, and a summary of the jresults with the most recent performance
standard test. Any observed abnormalities in reproductive behavior or performance of
control fish also must be reported.
3. Test Conditions
The report must specify the conditions underj which the test was performed. This must
include information on the source, treatment of, and basic chemical characteristics of the
i
dilution water. It also must include mean (± SD) and range for water temperature, dissolved
oxygen, pH, hardness, alkalinity, total organic carbon, and un-ionized ammonia. The
photoperiod and light intensity used during the exposure must be specified. The chamber
size, water volume, flow rate, and number and composition of spawning substrates must be
included. The general experimental design must be described including the number of
treatments, number of replicates (test chambers) per treatment, and the number of males and
females per test chamber. The report must include information on the feeding regime,
i
\
including priority contaminant levels, food supplier and lot number. The basic nature of the
exposure (i.e., flow-through, i.p. injection, dietary) must be stated, in addition to specific
86
-------
information related to the exposure type. For example, if the exposure was a flow-through
water delivery type, the daily number of volume exchanges of dilution water must be stated.
If a solvent or dispersant was used to deliver the chemical, the specific solvent or dispersant
and the concentrations to which the fish were exposed must be specified. If the exposure
route was by i.p. injection, the carrier and injection volume must be stated. If the exposure
route was dietary, the food items used to introduce the chemical must be stated in the report.
The methods used for, and results (with standard deviations) of the test chemical analyses
should be reported, including validation studies, and method detection/quantitation limits.
4. Results
The summary report must include the mean and standard deviation for each test endpoint.
Qualitative data collected on reproductive behavior and gonadal histopathology endpoints
must be summarized and reported. The results must include data for the control (plus solvent
control or sham i.p.-injected control, when used) and the treatment fish. Statistical
significance, as well as the statistical tests used, level of significance, and sample size must
be indicated.
K. Quality Assurance
Each laboratory should incorporate a strong quality assurance (QA) program from the start of
testing. Prior to the actual initiation of the pre-exposure period, Standard Operating Procedures
87
-------
(SOPs) should be prepared for all aspects of the study, including culturing of test animals,
operation of the test system, generation of stock solutions, instrument calibration, analysis of test
chemical, analyses of water quality characteristics, sample tracking and chain-of-custody
procedures, performance of pre-exposure test, an i performance of the 21-d reproduction test.
Staff that will be performing tasks on the project
tests begin.
should be familiar with the SOPs before the
QA practices must address all activities that affect the quality of the final data obtained from the
test, such as: (1) test substance, (2) source and condition of the test animals, (3) water quality
characteristics, (4) condition of exposure apparatus and equipment, (5) maintenance of chemical
exposure concentrations, (6) analytical methodology, (7) instrument calibration, (8) replication,
(9) record keeping, and (10) data evaluation and
interpretation. During routine activities,
established quality control (QC) practices should
that are of known quality.
be followed which ensure generation of data
The fish used in the test should appear healthy, b shave normally, feed well, reproduce
successfully, and have low mortality in culture,.pre-exposure period, and in the test controls.
i
Routine water quality characteristics such as temperature, dissolved oxygen, pH, hardness,
i
alkalinity, total organic carbon, and un-ionized ammonia should be monitored to assure quality of
the system used to maintain the animals. All instruments should be calibrated and standardized
following the instrument manufacturer's procedures, and monitored according to standard
methods. I
88
-------
Proper record keeping is important. A complete file should be maintained for each test
indicating test chemical, investigator, and dates of initiation and termination of test. The file
should contain the original data sheets from the test; source and information about the test
chemical and concentration of any solvent used; detailed records of the animals used in the test,
such as source, age, and any other pertinent information; test conditions employed; description of
the experimental design; methods used for, and results of, analysis of the test chemical; source
and characteristics of dilution water; information on the calibration of equipment and
instruments; and the methods used for, and results of, statistical analyses of the data. The file
should also contain anything unusual about the test, any deviations from established SOPs, and
all other relevant information. Laboratory data should be recorded on a real-time basis to prevent
loss of information or inadvertent introduction of errors into the record. Original data sheets
should be signed and dated by the laboratory personnel performing the tests.
Internal audits should be performed by the testing laboratory's QA officer during the actual
performance of each test by inspection of the SOPs, evaluation of the performance of the test
system, and inspection of the biological and chemical record books. Post-test audits should be
performed of the data analysis, statistical treatment, and interpretation of results, with records
made of any deviations from the test protocols. QA records should address test animal health,
pre-exposure reproductive performance, the performance of the test from a biological
perspective, and the performance of the test from a chemical perspective. A QA summary should
be prepared by the QA officer for each test and should accompany the test report.
89
-------
L. Interpretation of Results
The specific test described in this document, as Well as relatively similar reproduction studies
with the fathead minnow (e.g., Kramer et al 1998; Giesy et al 2000; Harries et al. 2000) have
been conducted with a number of EDCs representative of several MOA. In this1 section we
i
provide an overview of responses in the test to chemicals with known endocrine MOA, as a basis
for interpretation of study results with unknown chemicals. Some individual responses are very
i
diagnostic in terms of identification of a specific endocrine MOA (e.g., induction of vitellogenin
[
in males caused by estrogen receptor agonists), bijit in many cases it is/will prove necessary to
consider patterns of responses in the whole suite of endpoints to assess which (if any) endocrine
i.
pathway has been affected. It must be noted that (he database from which this interpretive
guidance was developed is limited. For .example, tests with chemicals with mixed (endocrine)
MOA have been rare, and likely would result in unanticipated patterns of responses (e.g., see
methyltestosterone example below). Another important shortcoming in the current knowledge
base is a lack of data for chemicals which affect reproduction, but not through alterations in the
I
endocrine systems of concern. The assumption in these cases is that some generic measure of
reproductive potential would be affected (e.g., fecundity, GST) in the absence of changes in other,
!
more diagnostic, endpoints such as secondary sex! characteristics, plasma vitellogenin and sex
steroid concentrations, and gonadal histopathology.
Table 5 summarizes responses of fathead minnows to different EDCs in the context of the suite
of endpoints described in this document. The most work, by far, has been with estrogen receptor
90
-------
agonists. Strong agonists, such as p-estradiol, reduce fecundity of actively-spawning animals,
and consistently induce vitellogenin in males (Table 5; Kramer et al. 1998; Panter et al. 1998;
2002; Tyler et al 1999; Korte ei al. 2000). Other endpoints that have been reported to be
affected by strong estrogen receptor agonists in sexually-mature fathead minnows include
gonadal (testicular and ovarian) histopathology and alterations in secondary sex characteristics
(Panter et al 1998; Miles-Richardson et al 1999a; Harries et al 2000). Exposure of fathead
minnows to chemicals that are weaker estrogen receptor agonists (e.g., alkylphenols,
methoxychlor) elicit a qualitatively similar pattern of effects similar to those observed after
exposure to stronger agonists, although the magnitude of the effects (not surprisingly) differs
between weak and strong estrogens (Miles-Richardson et al 1999b; Giesy et al 2000; Harries et
al. 2000; Ankley et al 2001). For example, methoxychlor significantly decreased (but did not
completely inhibit) spawning of fathead minnows at a concentration of about 5 jxg/L (Ankley et
al 2001). At this concentration, a significant induction of vitellogenin in male fathead minnows
was observed; however, the response was much less pronounced than when adult male fathead
minnows were exposed to strongly estrogenic substances (Panter et al. 1998; Korte et al. 2000;
Ankley et al 2001). There also have been descriptions of alterations in secondary sex
characteristics and ovarian histopathology in adult fathead minnows exposed to weak estrogens
(Miles-Richardson et al 1999b; Harries et al 2000; Ankley et al 2001). Plasma concentrations
of sex steroids also can be affected (in a sex-specific manner) by weak estrogen receptor agonists
(Table 5; Ankley et al. 2001); presumably, if comparable data were available, this also would be
observed in exposures with strong estrogens.
91
-------
One androgen receptor agonist, the synthetic compound methyltestosterone, has been evaluated
E
I
using the short-term fathead minnow reproduction test (Table 5; Ankley et al. 2001). It appeared
that the methyltestosterone elicited a suite of responses indicative of a chemical with a mixed
i
estrogenic and androgenic MOA rather than a "pure" androgen. Exposure to methyltestosterone
at concentrations > 0.2 mg/L caused an immediate cessation of spawning. Consistent with
previous demonstrations (from aquaculture studies) that methyltestosterone is androgenic in fish,
the adult females were clearly masculinized, exhibiting pronounced nuptial tubercle development
I
within about 6 d of exposure. However, methyltestosterone also caused a large induction of
vitellogenin in both males and females, which is a response consistent with (and relatively
specific to) an estrogen receptor agonist. This may have occurred because methyltestosterone
can be converted via aromatase to a methyl-estradiol analogue (Dr. D. Kime, University of
.k.
Sheffield, personal communication), which would have resulted in the fish actually being
exposed to an estrogen/androgen mixture. Given this, it is difficult to say whether other
responses observed in the test (e.g., reduced steroid concentrations, reduced GSI, abnormal
i
gonadal histology; Table 5) were due to the androgenic or estrogenic (or combined) nature of
methyltestosterone. Based, however, upon this siudy and early work by Smith (1974) with
fathead minnows exposed to known androgens, masculinization of adult females would appiear to
i
be a very diagnostic response for this MOA. i
i
j
i
1
Two putative anti-androgens have been evaluated in 21-d reproduction studies with the fathead
minnow (Table 5). Makynen et al. (2000) assessed the effects of vinclozolin on fathead
I
minnows held in a paired-breeding situation. Due to minimal reproduction in controls from that
92
-------
experiment, it was difficult to determine whether exposure to the mammalian androgen receptor
antagonist affected fecundity of the fish. Vinclozolin did not markedly affect plasma steroid
concentrations in males or females, and vitellogenin was not measured in that experiment. At a
concentration of about 700 u-g/L vinclozolin did, however, cause a significant reduction in GSI of
the females, which was accompanied by retarded obcyte maturation and atresia. Because neither
vincldzolin or its primary metabolites bound to the fathead minnow androgen receptor in vitro
(Makynen etal 2000), it was uncertain whether the responses observed in the gonads of the
females were truly indicative of an anti-androgen. Therefore, the results of a reproduction test
with the androgen receptor antagonist flutamide (which does bind to the fathead minnow
androgen receptor; Makynen et al. 2000) may be more descriptive of the expected pattern of
responses associated with exposure of reproductively-active fathead minnows to an anti-
androgen. At concentrations ranging from 60 to 600 ng/L, flutamide caused a concentration-
dependent decrease in fecundity which, as was observed with vinclozolin, was accompanied by
'• , -
decreased GSI and increased oocyte atresia/retarded maturation in the female fathead minnow
(Table 5; Jensen et al. 2002). Flutamide also affected steroid concentrations in both sexes and
appeared to cause a slight increase in vitellogenin concentrations in the female. In addition,
flutamide exposure resulted in subtle indications of gonadal histopathology in the males, which
was comprised of germ cell necrosis and reduced spermatogenesis. Based on these results, the
most consistent effects of the anti-androgens in this test appear to be expressed in the gonads of
the females (GSI, histopathology).
93
-------
Aromatase (CYP19) is a cytochrome P450-based enzyme that, under normal physiological
conditions, converts testosterone to P-estradiol. There is emerging evidence that the MOA via
which some EDCs exert their effects is through alterations in steroid synthesis associated with
inhibition of aromatase activity (i.e., CYP19). Fadrozole, a classical inhibitor of aromatase
i
activity, was evaluated using the protocol described in this document (Table 5; Ankley et al.
2002). The chemical caused a concentration-dependent reduction in fecundity at concentrations
ranging from about 1.5 to 50 jig/L. Consistent w^th the presumed MOA, there also was a
concentration-dependent decrease in both plasma P-estradiol and vitellogenin in the female
fathead minnow. In addition, plasma concentrations of testosterone and 11-ketotestosterone were
increased in the males, and histological alterations observed in the gonads of both sexes. Given
the specificity of aromatase inhibitors, the decreases in P-estradiol and, subsequently,
vitellogenin in the female fathead minnow should be an excellent diagnostic response for mis
class of EDCs. Previous studies with fish have emphasized vitellogenin induction in males as a
i
highly-specific indicator of an endocrine MOA (estrogen receptor agonists); these data indicate
an equally useful and diagnostic response associated with vitellogenin reductions in (sexually-
mature) females. This endpoint presumably would reflect effects of chemicals, not only on P-
estradiol synthesis (as for fadrozole), but the action of chemicals that act as estrogen receptor
antagonists (Panter et al. 2002)
The patterns of responses summarized in Table 5! clearly represent only a small subset of possible
f
outcomes associated with exposure to EDCs. Adverse effects associated with some MO As
t
(estrogen receptor agonists, androgen receptor agonists, and aromatase inhibitors) should be
94
-------
easily identified. Identification of chemicals as anti-ahdrogens may be more equivocal (although
these types of chemicals would clearly be "flagged" as endocrine-active through alterations in
gonadal histology and, perhaps, steroid concentrations). As this test is conducted with additional
chemicals reflective of the MOA discussed above, as well as other MO A, guidance in
interpreting test results will expand.
95
-------
CO
1
'§
CD
JS
O
bO
•a
f
CO
1
.c
•i
§
O
M
I
c
•s
0
^G
13
1
?
CD
U
g
o
(U
i
1
^
35
O
$25
H
IB
|
jj
w
"3
1
5
1
fl
*H
Q\
euo
1
K^.
H ^
1 S
S CC
s|
E? S
03 •**
1 §
1
£•
1
U
Em
MOA/Cheniica!
§>
§ a
45 0
0 'S
•a S
2 a
co S
8 ^
a) O
n3 Q
8 -8
•a -a
OT M
b «
1 2
O« 'co
S JJ
CU s
O w
O P>»
1 1
- §
*— *—
A
i §
*E3
U u
a 8
1 *
— *
^> *^a
Nu Q
^ S
•^ k,
B
ER Agonist
P-Estradiol a
S
g
4>
1
«~
S
H
~*
g
55
1
t
O
Methoxychlor
•55
fi
rt
H
0
0)
1
o
1
5
1
si
•1
I
8
•J3
I -
I
(L)
1 1
H
«
I I
1
H
H
1
AR Agonist
Methyltestost
AR Antagonist
Vinclozolin d
-------
cs
if
ON
ON
ON
§
CO
CS
• n.
oo
00 ,-t
ml
*O *^ ^3
11 g
•a -a -a
*'?-8?
H ^
00
97
-------
M. References
American Society for Testing and Materials. 2000a
toxicity tests with fishes, macroinvertebrates, and
Book of ASTM Standards. Volume 11.05. West
. Standard guide for conducting acute
amphibians. Method E 729-96. In Annual
Conshohocken, PA, USA, pp 213-233.
American Society for Testing and Materials. 20DOb. Standard guide for conducting early life-
stage toxicity tests with fishes. Method E 1241-98. In Annual Book of ASTM Standards.
Volume 11.05. West Conshohocken, PA, USA, pp 941-967.
Andreassen TK, Korsgaard B. 2000. Characterization of a cytosolic estrogen receptor and its
up-regulation by 17p-estradiol and the xenoestrogen 4-terf-octylphenol in the liver of the eelpout
(Zoarces viviparus). Comp Biochem Phys C 125:299-313.
Ankley GT, Giesy JP. 1998. Endocrine disrupters in wildlife: A weight of evidence
perspective. In Kendall R, Dickerson R, Suk W, Giesy J, eds, Principles and Processes for
Assessing Endocrine Disruption in Wildlife. SETAC, Pensacola, EL, USA, pp 349-368.
Ankley GT, Johnson RD, Detenbeck NE, Bradbury SP, Folmar L, Toth G. 1997. Development
of a research strategy for assessing the ecological risk of endocrine disrupters^ Rev Toxicol 1:71-
106.
Ankley GT, et al. 1998a. Overview of a workshop on screening methods for detecting potential
(anti-) estrogenic/androgenic chemicals in wildlife. Environ Toxicol Chem 17:68t87.
Ankley GT, Tietge JE, DeFoe DL, Jensen KM, Holcombe GW, Durhan EJ, Diamond SA.
1998b. Effects of ultraviolet light and methoprene on survival and development of Rana pipiens.
Environ Toxicol Chem 17:2530-2542. |
Ankley GT, Jensen KM, Kahl MD, Korte JJ, Makynen EA. 2001. Description and evaluation of
a short-term reproduction test with the fathead minnow (Pimephales promelas). Environ Toxicol
Chem 20:1276-1290. j
Ankley GT, Kahl MD, Jensen KM, Homung MW, Korte JJ, Makynen EA, Leino RL. 2002.
Evaluation of the aromatase inhibitor fadrozole in a short-term reproduction assay with the
fathead minnow (Pimephales promelas). Toxicol Sci 67:121-130.
Bardach JE, Bernstein JJ, Hart JS, Brett JR. 1966. Tolerance to temperature extremes: Animals.
Part IV. Fishes. Li Altman PL, Dittmer D, eds, Environmental Biology. Fed. Am. Soc. Exp.
Biol., Bethesda, MD, USA, pp 37-80.
Benoit DA, Mattson VR, Olson DL. 1982. A continuous-flow mini-diluter system for toxicity
testing. Water Res 16:457-464.
98
-------
Box GEP, Hunter WG, Hunter JS. 1978. Statistics for Experimenters. John Wiley, New York,
NY, USA, 653 pp.
Bradbury SP. 1994. Predicting modes of toxic action from chemical structure: An overview.
SAR QSAR Environ Res 2:89-104.
Bradbury SP. 1995. Quantitative structure-activity relationships and ecological risk assessment:
An overview of predictive aquatic toxicology research, Toxicol Lett 79:229-237.
Bradbury SP, Lipnick RL. 1990. Introduction: Structural properties for determining mechanisms
of toxic action. Environ Health Perspect 87:181-182.
Brooke LT, Call DJ, Geiger DL, Northcott CE,eds. 1984. Acute Toxicities of Organic
Chemicals to Fathead Minnows (Pimephales promelas), Vol I. Center for Lake Superior
Environmental Studies, University of Wisconsin, Superior, WI, USA, 414 pp.
Brungs WA. 1971. Chronic effects of constant elevated temperature on the fathead minnow
(Pimephales promelas Rafinesque). trans Am Fish Soc 100:659-664.
Call DJ, Geiger DL, eds. 1992. Subchronic Toxicities of Industrial and Agricultural Chemicals
to Fathead Minnows (Pimephales promelas)^ Vol I. Center for Lake Superior Environmental
Studies, University of Wisconsin, Superior, WI, USA.
Call DJ, Brooke LT, Knuth ML, Poirier SH, Hoglund MD. 1985. Fish subchronic toxicity
prediction model for industrial organic chemicals that produce narcosis. Environ Toxicol Chem
4:335-341.
Call DJ, Markee TP, Geiger DL, Brooke LT, VandeVenter FA, Cox DA, Genisot KI, Robillard
KA, Gorsuch JW, Parkerton TF, Reiley MC, Ankley GT, Mount DR. 2001. An assessment of
the toxicity of phthalate esters to freshwater benthos. I. Aqueous exposures. Environ Toxicol
Chem 20:1798-1804.
Chang CF, Hung CY, Chiang MC, Lan SC. 1999. The concentrations of plasma sex steroids
and gonadal aromatase during controlled sex differentiation in grey mullet, Mugil cephalus.
Aquaculture 177:37-45.
Colborn T, Dumanoski D, Meyers JP. 1996. Our Stolen Future: Are We Threatening Our
Fertility, Intelligence, and Survival? A Scientific Detective Story. Dutton, New York, NY, USA.
Crim LW, Glebe BD. 1990. Reproduction. mSchreckCB,MoylePB,eds,Metfio^/0rF«/i
Biology. American Fisheries Society, Bethesda, MD, USA, pp 529-553.
99
-------
Crisp TM, et al. 1997. Special report on environmental endocrine disruption: An effects
assessment and analysis. EPA/630/R-96/012. Risk Assessment Forum, U.S. Environmental
Protection Agency, Washington, DC, USA. !
I
Daniel WW. 1978. Applied Nonparamentric Statistics. Houghton Mifflin Co., Boston, MA,
USA. 503 pp. [
[
DeFoeDL. 1975. Multichannel toxicant injection system for flow-through bioassays. JFish
Res Board Can 32:544-546. \
DeFoe D, Holcombe G. 1997. An innovative computer-driven, electronic diluter for flow-
through toxicity tests. Abstracts, 18th Annual Meeting, Society of Environmental Toxicology
and Chemistry, San Francisco, CA, USA, p 305.
Devine G. 1968. A study of the smallmouth bass in ponds with special consideration of
minnows and decapods as forage. MA Thesis, University of Missouri, Columbia, MO, USA,
115 pp.
DeVito M, et al. 1999. Screening methods for jthyroid hormone disrupters. Environ Health
Perspect 107:407-415.
Elonen GE, Spehar, RL, Holcombe GW, Johnscin RD, Fernandez JD, Erickson RJ, Tietge JE,
Cook PM. 1998. Comparative toxicity of 2,3,7,8-tetrachloro-p-dioxin to seven freshwater fish
species during early life-stage development. Environ Toxicol Chem 17:472-483.
Erickson RJ, Ankley GT, DeFoe DL, Kosian PA, Makynen EA. 1999. Additive toxicity of
binary mixtures of phototoxic polycyclic aromatic hydrocarbons to the oligochaete Lumbriculus
variegatus. Toxicol Appl Pharmacol 154:97-105.
Fernandez JD, Denny JS, Tietge JE. 1998. A simple apparatus for administering 2,3,7,8-
tetrachlorodibenzo-/?-dioxin to commercially available pelletized fish food. Environ Toxicol
Chem 17:2058-2062.
Flickinger SA. 1969. Determination of sexes in the fathead minnow. Trans Amer Fish Soc
98:526-527.
Gale WF, Buynak GL. 1982. Fecundity and spawning frequency of the fathead minnow - A
fractional spawner. Trans Am Fish Soc 111 :35-40.
Gast MH, Brungs WA. 1973. A procedure for separating eggs of the fathead minnow. Prog
Fish-Cult 35:54.
100
-------
Geiger DL, Northcott CE, Call DJ, Brooke LT, eds. 1985. Acute Toxicities of Organic
Chemicals to Fathead Minnows (Pimephales promelas), Vol n. Center for Lake Superior
Environmental Studies, University of Wisconsin, Superior, WI, USA, 326 pp.
Geiger DL, Poirier SH, Brooke LT, Call DJ, eds. 1986. Acute Toxicities of Organic Chemicals
to Fathead Minnows (Pimephales promelas), Vol EL Center for Lake Superior Environmental
Studies, University of Wisconsin, Superior, WI, USA, 328 pp.
Geiger DL, Call DJ, Brooke LT, eds. 1988. Acute Toxicities of Organic Chemicals to Fathead
Minnows (Pimephales promelas), Vol IV. Center for Lake Superior Environmental Studies,
University of Wisconsin, Superior, WI, USA, 355 pp.
Geiger DL, Brooke LT, Call DJ, eds. 1990. Acute Toxicities of Organic Chemicals to Fathead
Minnows (Pimephales promelas), Vol V. Center for Lake Superior Environmental Studies,
University of Wisconsin, Superior, WI, USA, 332 pp.
Gelinas D, Pitoc GA, Callard GV. 1998. Isolation of a goldfish brain cytochrome P450
aromatase cDNA: mRNA expression during the seasonal cycle and after steroid treatment. Mol
Cell Endocrinol 138:81-93.
Giesy JP, Pierens SL, Snyder, EM, Miles-Richardson S, Kramer VJ, Snyder SA, Nichols KM,
Villeneuve DA. 2000. Effects of 4-nonylphenol on fecundity and biomarkers of estrogenicity in
fathead minnows (Pimephales promelas). Environ Toxicol Chem 19:1368-1377.
Gray LE, et al. 1997. Endocrine screening methods workshop report: Detection of estrogenic
and androgenic hormonal and antihormonal activity for chemicals that act via receptor and
steroidogenic enzyme mechanisms. Reprod Toxicol 11:719-750.
Harries JE, Runnalls T, Hill E, Harris CA, Maddix S, Sumpter JP, Tyler CR. 2000.
Development of a reproductive performance test for endocrine disrupting chemicals using pair-
breeding fathead minnows (Pimephales promelas). Environ Sci Technol 34:3003-3011.
Held JW, Peterka JJ. 1974. Age, growth and food habits of the fathead minnow, Pimephales
promelas, in North Dakota saline lakes. Trans Am Fish Soc 103:743-756.
Hoke RA, Ankley GT, Cotter AM, Goldenstein T, Kosian PA, Phipps GL, Vandermeiden FM.
1994. Evaluation of equilibrium partitioning theory for predicting acute toxicity of field-
collected sediments contaminated with DDT, DDE and DDD to the amphipod Hyalella azteca.
Environ Toxicol Chem 13:157-166.
Hoke RA, Kosian PA, Ankley GT, Cotter AM, Vandermeiden FM, Phipps GL, Durban EJ.
1995. Check studies with Hyalella azteca and Chironomus tentans in support of the
development of a sediment quality criterion for dieldrin. Environ Toxicol Chem 14:435-443.
101
-------
Jarvinen AW, Ankley GT. 1999. Linkage of Effects to Tissue Residues: Development of a
Comprehensive Database for Aquatic Organisms Exposed to Inorganic and Organic Chemicals.
SETAC, Pensacola, FL, USA, 364 pp.
Jensen KM, Korte JJ, Kahl MD, Pasha MS, Ankley GT. 2001. Aspects of basic reproductive
biology and endocrinology in the fathead minnow (Pimephales promelas). Comp Biochem
PhysiolC 128:127-141.
Jensen KM, Kahl MD, Korte JJ, Makynen EA, Leino RL, Ankley GT. 2002. Effects of
fiutamide in a short-term reproduction assay witn the fathead minnow (Pimephales promelas). In
preparation. ;
Kahl MD, Makynen EA, Kosian PA, Ankley GT. 1997. Toxicity of 4-nonylphenol in a life-
cycle test with the midge Chironomus tentans. Ecotoxicol Environ Sa/38:155-160.
Kahl MD, Russom CL, DeFoe DL, Hammermeister DE. 1999. Saturation units for use in
aquatic bioassays. Chemosphere 39:539-551. |
i
Kahl MD, Jensen KM, Korte JJ, Ankley GT. 2001. Effects of handling on endocrinology and
reproductive performance of the fathead minnow. JFishBiol (in press).
i
I
i
Kavlock RJ, et al. 1996. Research needs for the risk assessment of health and environmental
effects of endocrine disrupters: A report of the U.S. EPA-sponsored workshop. Environ Health
Perspect 104:715-740. !
Keith JJH. 1997. Environmental Endocrine Disruptors: A Handbook of Property Data. John
Wiley & Sons, New York, NY, USA. !
\
Kenaga EE. 1982. Predictability of chronic toxicity from acute toxicity of chemicals in fish and
aquatic invertebrates. Environ Toxicol Chem 1:347-358.
Kime DE. 1998. Endocrine Disruption in Fishl Kluwer Academic Publishers, Norwell, MA,
USA, 396 pp.
t
Kitano T, Takamune K, Nagahama Y, Abe SI. 2000. Aromatase inhibitor and 17a-
methyltestosterone cause sex reversal from genetical females to phenotypic males and
suppression of P450 aromatase gene expression in Japanese flounder (Paralichthys olivaceus).
Mol Reprod Develop 56:1-5.
Korte JJ, Kahl MD, Jensen KM, Pasha MS, Parks LG, LeBlanc GA, Ankley GT. 2000. Fathead
minnow vitellogenin: complementary DNA sequence and messenger RNA and protein
expression after 17p-estradiol treatment. Environ Toxicol Chem 19:972-981.
102
-------
Kramer VJ, Miles-Richardson S, Pierens SL, Giesy JP. 1998. Reproductive impairment and
induction of alkaline-labile phosphate, a biomarker of estrogen exposure, in fathead minnows
(Pimephales promelas) exposed to waterborne 17p-estradiol. Aquai Toxicol 40:335-360.
Lange R, Hutchinson TH, Croudace CP, Siegmund F, Schweinfurth H, Hampe P, Panter GH,
Sumpter JP. 2001. Effects of the synthetic estrogen 17cc-ethinylestradiol on the life-cycle of the
fathead minnow (Pimephales promelas). Environ Toxicol Chem 20:1216-1227.
Lewis PA, Klemm DJ, Lazorchak JM, Norberg-King TJ, Peltier WH, Heber MA. 1994. Short-
term methods for estimating the chronic toxicity of effluents and receiving waters to freshwater
organisms. 3rd ed. EPA/600/4-91/002. U.S. Environmental Protection Agency, Environmental
Monitoring Systems Laboratory, Cincinnati, OH.
Lewis RJ,Sr. 1991. Reproductively Active Chemicals. A Reference Guide. VanNostrand
Reinhold, New York, NY, USA.
Makynen EA, Kahl MD, Jensen KM, Tietge JE, Wells KL, Van Der Kraak G, Ankley GT. 2000.
Effects of the mammalian antiandrogen vinclozolin on development and reproduction of the
fathead minnow (Pimephales promelas). Aquat Toxicol 48:461-475.
Mayer FL, Jr, Ellersieck MR. 1986. Manual of acute toxicity: Interpretation and data base for
410 chemicals and 66 species of freshwater animals. Resource Publication No. 160. U.S. Fish
and Wildlife Service, Washington, DC, USA, 506 pp.
McCarraher DB, Thomas R. 1968. Some ecological observations on the fathead minnow,
Pimephales promelas, in the alkaline waters of Nebraska. Trans Am Fish Soc 97:52-55.
McKim JM. 1977. Evaluation of tests with early life stages of fish for predicting long-term
toxicity. / Fish Res Board Can 34:1148-1154.
Melo AC, Edmunds JSG, Ramsdell JS. 1999. Localization and sex steroid regulation of brain
aromatase activity: identification of a positive feedback mechanism that may be triggered by
environmental cues. In Henshel DS, Black MC, Harrass MC, eds, Environmental Toxicology
and Risk Assessment: Standardization of Biomarkers for Endocrine Disruption and
Environmental Assessment: 8h Volume, ASTM STP1364. American Society for Testing and
Materials, West Conshohocken, PA, USA, pp 226-238.
Miles-Richardson SR, Kramer VJ, Fitzgerald SD, Render JA, Yamini B, Barbee SJ, Giesy JP.
1999a. Effects of waterborne exposure of 17p-estradiol on secondary sex characteristics and
gonads of fathead minnows (Pimephales promelas). Aquat Toxicol 47:129-145.
Miles-Richardson SR, Pierens SL, Nichols KM, Kramer VJ, Snyder EM, Snyder S A, Render JA,
Fitzgerald SD, Giesy JP. 1999b. Effects of waterborne exposure to 4-nonylphenol and
103
-------
nonylphenol ethoxylate on secondary sex characteristics and gonads of fathead minnows
(Pimephales promelas). Environ Res 80:8122-8137.
Mommsen TP, Walsh PJ. 1988. Vitellogenesis
and oocyte assembly. In Hoar WS, Randall VJ
eds., Fish Physiology, Vol. XIA. Academic Press, San Diego, CA, USA, pp 347-406.
Mount DI. 1973. Chronic effect of low pH on fathead minnow survival, growth and
reproduction. Water Res 7:987-993.
I
Mount DI, Brungs WA. 1967. A simplified dosing apparatus for fish toxicology studies. Water
Res 1:21-29. [
Odeh RE. 1971. On Jonckherre's k-sample test against ordered alternatives. Technometrics
13:912-918. I
Pakdel F, F<5on Si Le Gac F, Le Menn F, Valotaire Y. 1991. In vivo estrogen induction of
hepatic estrogen receptor mRNA and correlation with vitellogenin mRNA in rainbow trout. Mol
CellEndocrinol75:205-212. !
f
i
Panter GH, Thompson RS, Sumpter JP. 1998. Adverse reproductive effects in male fathead
minnows (Pimephales promelas) exposed to environmentally relevant concentrations of the
natural oestrogens, oestradiol and oestrone. Aquat Toxicol 42:243-253.
Panter GH, Hutchinson TH, LUnge R, Lye CM, Sumpter JP, Zerulla M, Tyler CT. 2001. Utility
of a juvenile fathead minnow screening assay for detecting (anti)-estrogenic substances. Environ
Toxicol Chem (in press). !
Parks LG, Cheek AO, Denslow ND, Heppell 8A, McLachlan JA, LeBlanc GA, Sullivan CV.
1999. Fathead minnow (Pimephales promelas) vitellogenin: purification, characterization and
quantitative immunoassay for the detection of estrogenic compounds. Comp Biochem Physiol C
123:113-125.
i
Pelissero C, Lenczowski MJP, Chinzi D, DavailfCuisset B, Sumpter JP, Fostier A. 1996. Effects
of flavonoids on aromatase activity, an in vitro study. J Steroid Biochem Molec Biol 57:215-223.
Phipps GL, Holcombe GW. 1985. A method for aquatic multiple species toxicant testing: acute
toxicity of 10 chemicals to 5 vertebrates and 2 invertebrates. Environ Pollut (SerA) 38:141-157.
Pillard DA. 1995. Comparative toxicity of formulated glycol deicers and pure ethylene and
propylene glycol to Ceriodaphnia dubia and Pimephales promelas. Environ Toxicol Chem
14:311-315. !
104
-------
Poirier SH, Knuth ML, Anderson-Buchou CD, Brooke LT, Lima AR, Shubat PJ. 1986.
Comparative toxicity of methanol and N,N-dimethylformamide to freshwater fish and
invertebrates. Bull Environ Contam Toxicol 37:615-621.
Rawson DS, Moore JE. 1944. The saline lakes of Saskatchewan. Cancer Res 22:141-201.
Russom CL, Anderson EB, Greenwood BE, Pilli A, 1991. ASTER: an integration of the
AQUIRE data base and the QSAR system for use in ecological risk assessments. Sci Total
Environ 109/110:667-670.
Russom CL, Bradbury SP, Broderius SJ, Hammernieister DE, Drummond RA. 1997. Predicting
modes of toxic action from chemical structure: acute toxicity in the fathead minnow (Pimephales
promelas). Environ Toxicol Chem 16:948-967.
Scholz S, Gutzeit HO. 2000. 17a-ethinylestradiol affects reproduction, sexual differentiation
and aromatase gene expression of the medaka (Oryzias latipes). Aquat Toxicol 50:363-373.
\
Shilling AD, Carlson DB, Williams DE. 1999. Rainbow trout, Oncorhynchus mykiss, as a
model for aromatase inhibition. / Steroid Biochem Molec Bio 70:89-95.
Sharp JR. 1991. The influence of toxicants on the teratogenic response of fishes: a guide to the
literature, a glossary of terms and a pictorial atlas. Trace Substances in Environmental Health
24:63-80.
Smith, RJF. 1974. Effects of 17a-methyltestosterone on the dorsal pad and tubercles of fathead
minnows (Pimephales promelas). Can J Zool 52:1031-1038.
Smith, RJF. 1978. Seasonal changes in the histology of the gonads and dorsal skin of the
fathead minnow, Pimephales promelas. Can J Zool 56:2103-2109.
Society of Environmental Toxicology and Chemistry-Europe. 1997. Endocrine modulators and
wildlife: assessment and testing. In Tattersfield L, Matthiessen P, Campbell PM, Grandy N,
Lange R, eds. SETAC Europe/OECD/EC Expert Workshop on Endocrine Modulators, Wildlife-
Assessment and Testing. SETAC-Europe Press, Brussels, Belgium, 126 pp.
Sprague JB. 1969. Measurement of pollutant toxicity to fish. I. Bioassay methods for acute
toxicity. Water Res 3:793-821.
Sumpter JP, Jobling S. 1995. Vitellogenesis as a biomarker for estrogenic contamination of the
aquatic environment. Environ Health Perspect 103:173-178.
Suter GW H, Rosen AE, Linuer E, Parkhurst DF. 1987. Endpoints for responses of fish to
chronic toxic exposures. Environ Toxicol Chem 6:793-809.
105
-------
Thompson EA Jr, Siiteri PK. 1974. Utilization of oxygen and reduced nicotinamide adenine
dinucleotide phosphate by human placental niicrosomes during aromatization of
androstenedione. J Biol Chem 249:5364-5372.
Todo T, Dceuchi T, Kobayashi T, Nagahama Y. 1999. Fish androgen receptor: cDNA cloning,
steroid activation of transcription in transfected mammalian cells, and tissue mRNA levels.
Biochem Biophys Res Comm. 254:378-383. \
Tyler CR, van Aerie R, Hutchinson TH, Maddix S, Trip H. 1999. An in vivo testing system for
endocrine disrupters in fish early life stages using induction of vitellogenin. Environ Toxicol
Chem 18:337-347.
U.S. Environmental Protection Agency. 1976. Quality Criteria for Water (Red Book). Office
of Water Planning and Standards, Washington, DC, USA.
I
U.S. Environmental Protection Agency. 1982. IJser's guide for conducting life-cycle chronic
toxicity tests with fathead minnows. EPA/600/8-81-011. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Duluth, MN, USA.
U.S. Environmental Protection Agency. 1987. (3uidelines for the culture of fathead minnows
Pimephales promelas for use in toxicity tests. EPA/600/3-87/001. U.S. Environmental
Protection Agency, Environmental Research Laboratory, Duluth, MN, USA.
U.S. Environmental Protection Agency. 1993. Methods for Measuring the Acute Toxicity of
Effluents and Receiving Waters to Freshwater and Marine Organisms, 4th ed, Weber CI, ed.
EPA-600/4-90-027F. Office of Research and Development, Environmental Monitoring Systems
Laboratory, Cincinnati, OH, USA.
i
U.S. Environmental Protection Agency. 1994. Short-Term Methods for Estimating the Chronic
Toxicity of Effluents and Receiving Water to Freshwater Organisms, 3rd ed, Lewis PA, Klemm
DJ,LazorchakJM,Norberg-KingTJ, Peltier WH,Heber MA, eds. EPA-600-4-91-002. Office
of Research and Development, Environmental Monitoring Systems Laboratory, Cincinnati, OH,
USA.
U.S. Environmental Protection Agency. 1996a.
Development. EPA/600/R3-91-063. Washingto;
U.S. Environmental Protection Agency. 1996b.
minnow Pimephales promelas Rafinesque. EPA/600/R
Strategic Plan for the Office of Research and
n, DC, USA.
Prehatching development of the fathead
-96/079. Washington, DC, USA.
U.S. Environmental Protection Agency. 1998. Endocrine Disrupter Screening and Testing
Advisory Committee (EDSTAC) report. Office of Prevention, Pesticides and Toxic Substances,
Washington, DC, USA.
106
-------
Veith GD, Broderius SJ. 1990. Rules for distinguishing toxicants that cause Type I and Type
narcosis syndromes. Environ Health Perspect 87:207-211.
107
-------
APPENDIX A
DATA REPORTING FORMS
I
Form A-l. Record of Adult Fish Survival and Water Quality Characteristics
Form A-2. Record of Spawning Activity
Form A-3. Fathead Minnow Sample Record
Form A-4. Report Form for the 21-d Reproduction Test
108
-------
>H
•H
-S2
i
s
i
9
CO
CO
CO
CO
9
cs
s
3
s
g
u
T— 1
g
1
I— <
o
U
I
a
>
-------
O9
•a
i
o
•s
'I
-2
1
€ 2
«
£
£
I
9
en
-------
Treatment.
APPENDIX A-2
RECORD OF SPAWNING ACTIVITY
STUDY: _____^___
Replicate
Test
Day
1
2 .
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Init
Date
Total*
Spawns
'
#Eggs/:
Spawn
.•
1
# Fertile/
Spawn
(
t , '
!
Develop.
Stage
Incubation ID/
Comments
-
Ill
-------
APPENDIX A-3
FATHEAD MINNOW SAMPLE RECORD
STUDY:
Date:
Initials:
Treatment.
Replicate.
Sex
Body Weight
Gonad Weight
M / F
jng
Secondary Sex Characteristics/Comments
Ovipositor P/A
Tubercles P/A
Dorsal Pad P/A
Coloration P/A
Tissue Sample/ID Number/Comments
Plasma ID
Gonad
Histology ID
Aromatase ID
t
I Receptor ID
Liver ID
Brain ID
Body ID
Other comments:
Treatment
Replicate.
Sex
Body Weight
Gonad Weight
M / F
_mg
Secondary Sex Characteristics/Comments
Ovipositor P/A
Tubercles P/A
Dorsal Pad P/A
Coloration P/A
Tissue Sample/ID Number/Comments
Plasma ID
Gonad
Histology ID
Aromatase ID
Receptor ID
Liver ID
i
Brain ID
Body ID
Other comments:
112
-------
APPENDIX A-4
REPORT FORM FOR THE 21-D REPRODUCTION TEST
Chemical: _
Investigator:
Test Start Date:
Test Termination Date:
TEST CHEMICAL
CAS Number:
Chemical Source:
Lot Number:
Acute Toxicity:
Chronic Toxicity:
MF:
Kow: :
Water Solubility:.
Reference:
Reference:
MWT:
Vapor Pressure:
Purity:
Range-finding Test Information:
TEST ANIMALS
Source:
Condition:
Reference Toxicity Test
Chemical:
Results:
Date:
Age at Test Start:
Pre-Exposure Reproduction:
EDC Performance Standard Test
Chemical:
Results:
Date: .
TEST CONDITIONS
Water Temperature (°C) Mean (± SD)
Dissolved Oxveen (mc/L) Mean (± SD)
pH Mean (+ SD)
Wflrdne<5
-------
APPENDIX A-4 (Cont)
Type of Chemical Administration:.
(e.g. aqueous without solvent, aqueous with solvent, dietary, i.p. injection)
If aqueous administration was used, specify:
Method of Stock Generation: Mean Concentration of Stockfs):
If solvent carrier was used, specify:
Solvent: Maximum Solvent Concentration:
If dietary exposure was used, specify:
Food: Method of Chi
Source of Food: Food Lot Nurr
jmical Incorporation:
iber:
If i.p. exposure was used, specify:
Carrier Solvent: Carrier Solvent Concentration:
Volume of Carrier Solver
it Injected:
TEST CHEMICAL CONCENTRATIONS
Date
Nominal:
Control
Control8
t
Treatment 1
Treatment 2
Measured
Date:
Date:
Date:
Date:
Date:
Date:
Mean (± SD)
Range
Mean (± SD) % Recove
Mean (± SD) Repeatabil
Analytical Methodology
Method Detection/Quan
ry of Spiked Sampl<
ity of Duplicate An
ss: CN =
alvsis: CN = 1
)
l i
titation Limits
114
-------
APPENDIX A-4 (Cent)
TEST RESULTS (Mean ± SD)a
Primary Endpointsb
Adult Survival (%)
Reproductive Behavior
(specify)
Secondary Sex Characteristics
(specify ovipositor, tubercles,
dorsal pad, coloration)
GSI (%)
Gonadal Histopathology
(specify)
Plasma vitellogenin (rag/ml)
Plasma sex steroids (specify P-
estradiol, testosterone, 11-
ketotestosterone) (ng/ml)
Fecundity (specify total eggs,
number of spawns/female,
number of eggs/spawn)
Fertility (%)
Control
Control0
,v ; • '
Treatment 1
Optional Endpoints
Embryo Hatch (%)
Larval Survival (%)
Larval Morphology (specify)
Treatment 2
8 Statistical significance, as well as the statistical tests Used, level of significance, and sample size should
be indicated.
b Adult survival, reproductive behavior, secondary sex characteristics, GSI, gonadal histopathology,
plasma vitellogenin and sex steroids must be reported on a sex-specific basis.
0 Solvent control or sham-injected control.
General Remarks
115
-------
APPENDIX B
MEASUREMENT OF PLASMA VITELLOGENIN IN
FATHEAD MINNOWS BY COMPETITIVE ELISA
OPERATING PROCEDURE
1. SCOPE AND APPLICATION |
This procedure is used to at the MED laboratory to determine the concentration of
vitellogenin (Vtg) in plasma of fathead minnows (Pimephales promelas). Other validated
techniques can be used to measure Vtg; the following is included as one option. The level of
Vtg present is indicative of the presence of estradiol, or other estrogen-like compounds,
which can bind the estrogen receptor and induce the synthesis of Vtg. The ability to detect
Vtg in male fish is particularly significant since Vtg is normally undetectable or present at
extremely low levels compared to female fish. The Vtg is detected using antibodies in an
ELISA (enzyme linked immunosorbent assay) procedure. The detection limit for Vtg is
routinely 2 \ig per ml of plasma when using k 1:300 dilution, which is typical for most males.
2. SUMMARY OF METHOD
The assay is performed in 96-well EOS A plates that have been coated with purified fathead
minnow Vtg. After blocking the unbound sites with goat serum, the plate is washed to
remove any unbound material. The samples being assayed, a range of standards prepared
with purified fathead minnow Vtg, and the appropriate blanks and controls are incubated
with an antibody specific to fathead minnow! Vtg. After this incubation, the samples are
added to duplicate wells of the plate. The plate is incubated again, during which there is a
competition for the antibody between the Vtg bound on the surface of the wells and the Vtg
present in the samples and standards. This is followed by another wash which removes all
the primary antibody which is not bound to the Vtg on the plate. The primary antibody that is
bound to the plate is detected by incubation with a secondary antibody which is conjugated to
horseradish peroxidase. After washing the plate, the activity of the enzyme is measured,
which is inversely proportional to how much Vtg was present in the sample.
2.1 Definitions
Vtg = Vitellogenin
ELISA = Enzyme-linked Immunosorbent Assay
EHM = Fathead Minnow t
1° Ab = Primary Antibody
2° Ab = Secondary Antibody
NGS = Normal Goat Serum :
BSA = Bovine Serum Albumin
116
-------
IAP = Interassay Pool Sample
PBS = Phosphate Buffered Saline
PBST = Phosphate Buffered Saline with 0.05 % Tween-20 (same as wash buffer)
NSB = Nonspecific Binding
B0 = Maximum Binding
B = Binding
2.2 Health and Safety Warnings
. . -. f
Follow routine laboratory safety precautions such as wearing lab coats and safety
glasses. In addition, wear latex or nitrile gloves when working with phosphoric or
hydrochloric acid solutions to prevent severe damage by contact with skin.
2.3 Cautions
Vtg is a very labile protein; keep on ice and store at -20°C when not in use. Store
plasma samples and 1° Ab at -80°C. Store 2° Ab at -20°C. The 10X PBS buffer can be
stored for several months at 4°C; discard if the solution becomes cloudy.
3. PERSONNEL QUALIFICATIONS
Those who perform this assay should be trained and adept at performing sample dilutions and
working with microliter and multichannel pipets. Accurate pipeting is the most important
skill required to obtain quality results.
4. MATERIALS AND METHODS
Equipment:
Centrifuge for hematocrit tubes (e.g., Adams MHCT n)
Centrifuge for microcentrifuge tubes (e.g., Eppendorf 5417R)
Speed-Vac for lyophilizing aprotinin (e.g., Savant SVC 100)
Water bath at 37°C
Multi-tube vortexer (e.g., VWR)
Vortex mixer
96-well plate reader (e.g., Bio-Rad 3550)
Multichannel (8 or 12) pipet capable of delivering volumes in the 100 to 200 ul range
Adjustable pipets for 0.5 to 10 ul, 10 to 1000 ul, and 1 to 10 ml range
Repeater pipet for delivering 300 ul
Computer spreadsheet software (e.g., Lotus 123, version 9.5)
Computer graphing software (e.g., SlideWrite, version 5)
117
-------
Materials:
Heparinized microhematocrit tubes (e.g., Oxford 8889-301209)
0.6 ml microcentrifuge tubes
Aprotinin (e.g., Sigma A-6279)
Pipet tips
Parafilm
Air-tight container (e.g. Tupperware)
15 and 50 ml disposable centrifuge tubes
12 x 75 mm disposable glass test tubes
96 well EIA plates (e.g., ICN 76-381-04)
Purified FHM Vtg (Obtained from Dr. Nancy Denslow, University of Florida)
1° Ab for FHM Vtg (Obtained from Dr. Louise Parks, EPA)
Peroxidase-conjugated anti-rabbit 2° Ab (e.g., Bio-Rad 172-1019)
Normal goat serum (NGS) (e.g., Chemicon S26-100ml)
TMB peroxidase substrates (e.g., KPL 50-76-00)
1MH3P04
lAfHCl
Coating Buffer 50 mM carbonate buffer, pH 9.6:
1.26gNaHC03
0.68g NaaCOa
428 ml of deionized water ;
10X PBS (0.1 M phosphate, 1.5 MNaCl):
0.83g monobasic sodium phosphate (monohydrate)
20. Ig dibasic sodium phosphate (heptahydrate)
71gNaCl
810 ml of deionized water
Wash buffer (PBST):
100 ml of 10X PBS
900 ml of deionized water
0.5 ml Tween-20 (e.g., Sigma P-7949)
Adjust pH to 7.3 with 1M HC1
Blocking solution (5% NGS in Coating Buffer)
Assay buffer (PBST+ 2.5% NGS)
4.1 Sample Collection
Collect blood from severed caudal artery/vein with a heparinized microhematocrit tube
and place on ice or other cooling device. It is important to keep the blood cold to
minimize Vtg degradation. Centrifuge for 3 min, score tube and expel plasma into 0.6
ml microcentrifuge tubes containing 0.13 units of lyophilized aprotinin. (Prepare these
tubes in advance by adding the appropriate amount of aprotinin solution, freezing, and
118
-------
r
lyophilizing in a speed-vac at low heat for approximately 20 min, or until no liquid
remains). Mix samples gently and centrifuge briefly to collect the contents at the bottom
of the tube. Store at -80°C until analysis.
4.2 Analysis Procedure
The instructions describe the method for processing one plate at a time. Two or three
plates may be processed simultaneously, but standards and appropriate blanks/controls
must be included in each plate.
4.2.1 Coat Plate
Dilute purified FHM Vtg to 0.56 jig/ml in coating buffer. (The actual concentration
is determined empirically and may change with different batches of purified VTG).
Add 200 ul to each well of the plate. Cover the plate with parafilm and place in a
sealed container (e.g., Tupperware) along with some moistened paper towels.
Incubate the container overnight at 4°C (or for 2 h at 37°C).
4.2.2 Block Plate
Shake out the coating solution and pat the plate dry on absorbent paper. Add 350 ul
of blocking solution to each well, cover with parafilm, place in a container like above,
and incubate for 2 h at 37°C (or overnight at 4°C if the plate will not be used that day;
the plate should be used within a few days).
4.2,3 Prepare Standards
Dilute purified Vtg standard (concentration determined by supplier using the common
Bradford Method and comparison to BSA) to 2000 ng/ml and 750 ng/ml in 2 ml of
assay buffer. Mix one part (0.5 ml) of each of these with three parts (1.5 ml) of assay
buffer to yield 500 and 188 ng/ml concentrations, respectively. Prepare additional
dilutions in a similar manner until the entire range of desired standards is obtained
(2000, 750, 500, 188,125,47, 31,12, 8, and 3 ng/ml).
4.2.4 Dilute Samples
Because of normal variation in fathead minnow Vtg concentrations, the following
should be considered a general guideline. For a typical male with little or no Vtg,
combine 1 ul of plasma and 300 pi of assay buffer directly in the test tube that will be
used for incubation with the 1° Ab. For females or males that may have high Vtg
levels, additional dilution is usually required. Dilute an initial 300-fold dilution
further by combining 5 ul with 2 ml of assay buffer giving a total dilution of 120,000-
fold. One or two additional 1:2 dilutions (1 part diluted sample to 1 part assay buffer)
119
-------
of the 120K dilution are usually necessary to achieve a Vtg sample concentration
within the reliable range of the standards (approximately 10 to 200 ng/ml). If there
are enough sample wells available, it is advantageous to analyze samples of female
plasma (or male plasma if Vtg is highly elevated) at the 120K, 240K, and 480K
dilution, but just one of the dilutions is generally sufficient if the result is within the
reliable range of the standards. Put 300 ul of each dilution in a test tube for
incubation with the 1° Ab.
4.2.5 Prepare Interassay Pool (IAP) Sample
An interassay pool sample should be analyzed as a quality control procedure each
time an assay is performed. Sufficient plasma for this sample can be obtained by
combining unused plasma remaining from previously analyzed samples into a large
pool, separating into aliquots of a few ul, and storing at -80°C. Each time an assay is
performed, remove one of these aliquots and dilute as necessary to get a result in the
reliable range of the standards.
4.2.6 Prepare Maximum (B0) and Non-specific Binding (NSB) Samples
Prepare B0 by putting 300 ul of assay buffer in a test tube. Prepare NSB by putting
600 ul of assay buffer in a test tube.
4.2.7 Incubate with Primary Antibody
The 1° Ab must be diluted to give the desired response. The antibody has previously
been diluted 1:2 with an equal volume of normal male FHM plasma. On the day of
use, dilute the antibody UK-fold (1.8 ul to 20 ml) in assay buffer. This dilution is
determined empirically and will likely change from batch to batch. After diluting the
1° Ab, add 300 ul to each sample/standard tube (except the NSB sample) using the
repeater pipet. Mix the tubes briefly and incubate for 1 h at 37°C.
4.2.8 Wash Plate
I
Approximately 5 min before the end bf the 1° Ab incubation, shake out the blocking
solution and pat the plate dry on absorbent paper. Then, using a multichannel pipet,
fill the wells with 400 ul of wash buffer, remove by shaking, and pat dry. Repeat this
procedure two times. }
. i
4.2.9 Load Samples on Plate ,
i
At the end of the 1° Ab incubation, remove the tubes from the water bath and slightly
vortex. Remove two-200 ul aliquots and add to each of the duplicate wells of the 96-
well plate. A template identifying sample location is useful. After the samples are
120
-------
loaded, cover the plate with parafimi and incubate the container for 1 h at 37°C as
before.
4.2.10 Incubate with Secondary Antibody
Before the first use, divide the 2° Ab into aliquots of 5-10 ul and store at -20°C^
Fifteen min before the end of the sample incubation, remove an aliquot of the 2° Ab
from the freezer and dilute 40K (1 ul to 40 ml) with assay buffer. At the end of the
sample incubation, wash the plate as described above (4.2.8). Then add 200 ul of the
2° Ab solution to each well. Cover the plate with parafilm and incubate the container
for 1 h at 37°C as before.
4.2.11 Enzymatic Conversion of Substrate to Colored Compound
Approximately 15 min before the end of the 2° Ab incubation, turn on the plate
reader. Measure 7 ml each of the two peroxidase substrate solutions from the TMB
kit (this is sufficient for one 96-well plate) and allow solutions to come to room
temperature by placing them in a water bath. At the end of the 2° Ab incubation,
wash the plate as described above (4.2.8). Mix the two substrate solutions and add
100 ul of the mixture to each well of the plate using the multichannel pipet. Do this
in a timed manner. After 5 to 10 min, the color in the wells will change to blue. Stop
the reaction by adding 100 ul of 1 M phosphoric acid, which changes the color in the
wells to intense yellow. The exact time of reaction is not critical, but each well
should react for the same amount of time. The goal is to achieve a net absorbance of
1.5 to 2.0 in the wells with the most color.
4.2.12 Read Absorbance with Plate Reader
Set the plate reader to measure at 450 nm with the lower limit at 0.000 and the upper
limit at 3.000.
4.5 Troubleshooting
Normal results will produce a sigmoidal curve (see Section 5) with the steepest part
between 10 and 200 ng/ml. When changing to a new source of purified Vtg or 1° Ab,
the dilutions may have to be adjusted to obtain the proper curve. Lowering the
concentration of the coating Vtg will shift the curve to the left, or a more sensitive area.
Likewise, a more dilute 1° Ab will also make the assay more sensitive. While a more
sensitive assay is generally good, performing the assay in this range will narrow the
range of concentrations that can be detected. Due to the normal variation between fish,
having a narrow effective range of the standards would not necessarily be advantageous.
Also, making the assay more sensitive decreases the net absorbance response, which will
adversely affect detection.
121
-------
5. DATA ACQUISITION, CALCULATIONS, AND DATA REDUCTION
The absorbance values from the plate reader are obtained on a printout and entered into a
spreadsheet (e.g., Lotus 123 version 9.5) for performing the necessary calculations.
5.1 Calculate B/B0
Calculate the B/B0 for each sample and
standard. Divide the absorbance value by the
absorbance of the B0 sample, after the NSB value has been subtracted. Determine the
B/B0 for each duplicate well of the standards individually; for the samples, determine the
mean B/B0 of the two duplicate wells.
i,
5.2 Obtain Standard Curve
I
Generate the sigmoidal standard curve using graphing software (e.g. SlideWrite 5). Plot
the B/B0 of the standards against the amount of standard present on a log scale.
5.3 Determine Corresponding Amount of Vtg in Samples from B/B0 Values
Use the sample B/B0 values from the spreadsheet to generate extrapolated values of Vtg
present in the sample using the graphing software.
5.4 Calculate Amount of Vtg per ml of Plasma
Transfer the extrapolated sample Vtg values to the spreadsheet. Multiply these values by
the appropriate dilution factor and convert to mg/ml by dividing by 1,000,000, giving mg
Vtg per ml of plasma.
5.5 Determine Minimum Detection Limits
Often, particularly in normal males, it may be difficult to determine whether the result is
truly zero or slightly greater than zero. If this is the case, use the 95% confidence limit
(calculated with spreadsheet software) to determine if the result should be reported as
zero or as greater than zero. If the sample result is within the confidence interval of the
zero standard (NSB), report the result as' zero. The minimum detection level is the
lowest standard which is consistently different than the zero standard; that is, the two
confidence limits do not overlap. For any sample result which is within the confidence
limit of the minimum detection level or above, report the calculated value. If the sample
result falls between the zero standard and the minimum detection level, report one-half
of the minimum detection level.
122
-------
5.6 Data Management and Records Management
Attach printed versions of any computer generated data in a bound record book along
with any other necessary descriptions of how the data were generated.
6. QUALITY CONTROL AND QUALITY ASSURANCE SECTION
1 ' ' i
6.1 Sigmoidal Shaped Standard Curve
The standard curve should have a typical sigmoidal appearance, flattening out at both the
low and high end of the standard range. The coefficient of determination for the curve
(determined from graphing software) should be around 0.99. If it is sufficiently
different, the plot should be examined to determine if there are potential outliers. Also,
the sensitivity should not differ from previous analyses. This would be indicated by a
change in the steeper portion of the curve relative to the standards.
6.2 Low NSB Value
The NSB should be approximately 0.06, or less.
6.3 Consistent IAP Results.
The IAP result should not differ from previous analyses, or drift slowly over several
experiments.
6.4 Dependable Pipets
Check the operation of the pipets periodically for accuracy and precision. Fix or replace
any pipets not operating correctly.
6.5 Duplicate Agreement
Duplicate agreement should be within 10%.
6.6 Agreement with Different Dilutions
When multiple dilutions of the same sample are assayed, the final results should be
reasonably similar. However, the most accurate results are obtained when the B/B0
values are approximately 50%.
123
-------
7. REFERENCES i
j
Parks LG, Cheek AO, Denslow ND, Heppell SA, McLachlan JA, LeBlanc GA, Sullivan CV.
1999. Fathead minnow (Pimephales promelas) vitellogenin: purification, characterization,
and quantitative immunoassay for the detection of estrogenic compounds. Comp Biochem
Physiol €123:113-125.
Jensen KM, Korte JJ, Kahl MD, Pasha MS, Ankley GT. 2001. Aspects of basic
reproductive biology and endocrinology in the fathead minnow (Pimephales promelas).
Comp Biochem Physiol C 128: 127-141.
Korte JJ, Kahl MD, Jensen KM, Pasha MS,|Parks LG, LeBlanc GA, Ankley GT. 2000.
Fathead minnow vitellogenin: complementary DNA sequence and messenger RNA and
protein expression after 17p-estradiol treatment. Environ Toxicol Chem 19:972-981.
124
-------
APPENDIX C
DETERMINATION OF PLASMA STEROID
CONCENTRATIONS IN FATHEAD MINNOWS BY RIA
OPERATING PROCEDURE
1. SCOPE AND APPLICATION
This procedure is used at the MED laboratory to measure the concentration of sex steroids in
plasma of fathead minnows (Pimephales promelas) by radioimmunoassay (RIA). With
relatively minor changes in reagents, it is used to measure p-estradiol (E2), testosterone (T),
and 11-ketotestosterone (11-KT). Other validated techniques can be used to measure these
steroids; the following is included as one option. The levels of steroids determined from
these assays are used to assess the endocrine status of fish used in research on endocrine
disrupting chemicals (EDCs). Typically, detection limits are about 0.4 ng/ml of plasma when
performing the assay as described herein. :
2. SUMMARY OF METHOD
This assay is performed on ether-extracts of plasma to minimize complications from
interfering substances. The extract is incubated with a small amount of radioactive steroid
and an antibody which recognizes the steroid of interest. The unlabeled steroid present in the
sample and the radioactive steroid compete to bind to the antibody. The more steroid present
in the sample, the less radioactive steroid the antibody can bind. The addition of a charcoal
solution binds all of the unbound steroid. After centrifugation to remove the charcoal, the
radioactivity is measured in an aliquot of the supernatant fluid.
2.1 Definitions
RIA = Radioimmunoassay
E2 = p-Estradiol
T = Testosterone
11-KT= 11-Ketotestosterone
BSA = Bovine Serum Albumin
PBS = Phosphate Buffered Saline
NSB = Nonspecific Binding
B0 = Maximum Binding
B = Binding
3H = Tritium or Tritiated
IAP = Lnterassay Pool
125,
-------
2.2 Health and Safety Warnings j
2.2.1 General
i
Follow routine laboratory safety precautions such as wearing lab coats and safety
glasses. In addition, wear latex or nitrile gloves when working with sodium
hydroxide, hydrochloric acid, ethanol, and scintillation cocktail to prevent severe
damage by contact with skin.
2.2.2 Working with Radioactivity i
Small amounts of 3H are used in these assays. Wear gloves when handling 3H.
Training in safe handling of radioactive materials is required by the NRC (Nuclear
Regulatory Commission).
2.2.3 Ether
Ether is a very volatile, extremely flammable, toxic compound and should be used in
a fume hood as much as possible. Cover tubes with parafilm when removing them
from the hood. Wear gloves to prevent contact with skin.
'
2.2.4 Steroids
These steroids are hormones and thus have activity in humans. They also may
possess carcinogenic and mutagenic activities. It is important not to come in contact,
ingest, or breathe these compounds. Wear gloves to prevent contact with skin.
2.3 Cautions
I
Care should be taken not to spill radioactive solutions, especially when working with the
stock vials. Work with absorbent toweling, preferably with plastic bacldng, under all
work areas. If a spill should occur, it must be cleaned up and the spill area
decontaminated in accordance with NRp requirements.
3. PERSONNEL QUALIFICATIONS
I
Those who perform this assay should have completed a radiation safety course. Personnel
should also be trained and accomplished at making sample dilutions and working with
microliter volumes.
126
-------
4. MATERIALS AND METHODS
Equipment:
Scintillation counter (e.g., Packard 2500 TR)
Centrifuge (e.g., Jouan CR412)
Vortex mixer
Multi-tube vortexer (e.g., VWR)
Water bath
Pipets for 0.5 to 10 ul, 10 to 1000 ul, and 1 to 10 ml range
Repeater pipet for delivering 100 ul
Repeater diluter for removing aliquot/adding scintillation fluid (e.g., Brinkman)
Bottle-top dispenser for dispensing ether (e.g., Brinkman)
Computer spreadsheet software (e.g., Lotus 123, version 9.5)
Computer graphing software (e.g., SlideWrite, version 5)
Materials:
Heparinized microhematocrit tubes (e.g., Oxford 8889-301209)
0.6 ml microcentrifuge tubes
Parafilm
Disposable 12 X 75 mm borosilicate tubes
7 ml glass scintillation vials
Pipet tips
p-Estradiol (e.g., Sigma E-2758)
Testosterone (e.g., Sigma T-1500)
11-ketotestosterone (e.g., Sigma K-8250)
Tritiated E2 (e.g., Amersham TRK 587)
Tritiated T (e.g., Amersham TRK 921)
Tritiated 11-KT (e.g., Amersham TRQ 7903, Custom Synthesis)
Antibody (antisera) for E2 (e.g., Endocrine Sciences E26-47)
Antibody (antisera) for T (e.g., Endocrine Sciences T3-125)
Antibody (antisera) 11-KT (e.g., Biosense Laboratories)
Scintillation cocktail (e.g., Fisher Scintiverse SX18-4)
Ethyl ether
Parafilm ,
Ethanol :
0.1 M Phosphate Buffered Saline (PBS), pH 7.6
11.36gNa2HP04
3.12gNaH2PO4'2H2O
8.76gNaCl
Deionized water to ~950 ml,
Adjust pH to 7.6 with solid NaOH and 1 M NaOH
Deionized water tolOOO ml
127
-------
Assay buffer, or 0.01 MPBS, pH 7.4, with 1% BSA
100 ml of 0.1M PBS,' pH 7.6
900 ml of deionized water
7.9gNaCl
Adjust pH to 7.4 with 1N HC1
Add 10 g BSA (e.g., Sigma A-7888)
Dextran-coated charcoal solution
1.5 g Activated charcoal (e.g., Sigma C-5260)
0.15 g Dextran (e.g., Sigma D-4751)!
300 ml of 0.1 M PBS, pH 7.6
Mix at 4°C overnight before use. Keep at 4°C until just before using.
Steroid Stock Solutions
A) Dissolve 10.0 mg steroid in 10.0
ml of ethanol
B) Mix 100 ul of (A) with 900 ul of ethanol
C) Mix 10 ul of (B) with 990 ul of ethanol, aliquot at ~ 20 [il/vial and store at -80°C
4.1 Sample Collection
Collect blood from severed caudal artery/vein with a heparinized microhematocrit tube
and place on ice or other cooling device. It is important to keep the blood cold since
these samples will often be used for vitellogenin (Vtg) analyses also. Centrifuge for 3
min, then score tube and expel plasma into 0.6 ml microcentrifuge tubes containing 0.13
units of lyophilized aprotinin. (Prepare these tubes in advance by adding the appropriate
amount of aprotinin solution, freezing, and lyophilizing in a speed-vac at low heat for
approximately 20 min, or until no liquicl remains. If Vtg will not be analyzed, tubes
containing aprotinin are not necessary). Mix samples gently and centrifuge briefly to
collect the contents at the bottom of the
tube. Store at -80°C until analysis.
4.2 Ether Extraction and Reconstitution of Plasma
A convenient number of samples to extract is approximately 24. In addition, an
interassay pool sample should be analyzed as a quality control procedure each time an
assay is performed. (Sufficient plasma for this sample can be obtained by combining
unused plasma remaining from previously analyzed samples into a large pool, separating
into aliquots, and storing at -80°C). Do not prepare more RIA tubes for any one steroid
analysis than the centrifuge can accommodate. It is important that all of the tubes for a
particular steroid analysis be centrifuged at the same time.
4.2.1 Measure Volume of Plasma
Thaw plasma and, if possible, remote 6 ul for every determination that is to be made.
For example, in males it is often desirable to measure E2, T, and 11-KT, and since it
is desirable to perform duplicate analyses, it would be optimal to have 36 ul of
128
-------
plasma. Often it is not possible to collect 36 ul of plasma; if this is the case, priority
must be assigned to individual steroids for analyses. Record the volume of plasma
removed from each sample and placed in the test tube for extraction so that it can be
properly reconstituted.
4.2.2 Add PBS ,
Add 150 ul of 0.1 M PBS to each tube. This is essentially to give some volume to the
aqueous phase, although pH and ionic strength may also be important.
4.2.3 Add Ether
Use the bottle-top dispenser, if available, to deliver 1.5 ml of ether to each tube. Lay
a sheet of parafilm over all the tubes and vortex for 1 min using the multi-tube
vortexer. Adjust speed to highest possible without the ether contacting the parafilm.
If ether contacts parafilm, some waxes may be extracted which will interfere with
redissolving the steroids.
4.2.4 Place Samples in Freezer
Wait 1 min after vortexing to allow the phases to completely separate. Place the rack
in a -80°C freezer for 10 min to freeze the lower aqueous phase.
4.2.5 Decant Ether
Pour the upper ether layer into another clean, labeled test tube. Try to get as much of
the ether as possible, but work quickly enough that all the aqueous phases remain
frozen.
4.2.6 Repeat Extraction
Thaw the aqueous phase and repeat steps 4.2.3 to 4.2.5. Decant the ether from the
second extraction into the same tube as the first extraction.
4.2.7 Evaporate Ether
Leave the tubes in the fume hood overnight to evaporate all the ether.
4.2.8 Reconstitute Samples
Add 120 ul of assay buffer to each tube for each 6 ul-aliquot removed in 4.2.1. Lay a
sheet of parafilm over all the tubes and vortex for 1 min using the multi-tube
vortexer. (It may be necessary to manually reconstitute the samples by rinsing the
129
-------
walls of the extraction tube with assay buffer when the volume is too low for the
vortexer to wet the sides of the tube.) If there is solution clinging to the sides of the
tube, it may be necessary to centrifuge briefly to collect the contents at the bottom.
4.3 Incubate Samples with Antibody and Radioactive-tracer
4.3.1 Prepare Standards
Thaw an aliquot of Standard Solution C for each steroid that will be measured.
Prepare a 10 ng/ml standard; add 10 ^tl of Standard Solution C to 990 ul of assay
buffer in a test tube and vortex to mix thoroughly. Complete a series of 1:2 (1 part to
1 part) dilutions with assay buffer (ejg. 0.5 ml) to obtain standards of 5,2.5,1.25,
0.625,0.312,0.156,0.078,0.039,0.020, and 0.010 ng/ml.
4.3.2 Prepare Antibody Solution
The following should be considered a general guideline; specific concentrations may
change with different batches of antibodies. Initial experimentation with new batches
should attempt to determine the appropriate dilution resulting good sensitivity
between the 0.1 to 1.0 ng/ml standard range.
4.3.2.1 Estradiol
Resuspend as directed by the supplier, aliquot in approximately 100 ul portions,
and store at -80°C. On the day of use, thaw an aliquot and dilute 70 ul with 10 ml
of assay buffer. Mix gently, but thoroughly.
4.3.2.2 Testosterone
Resuspend as directed by the supplier, aliquot in approximately 100 ul portions,
and store at -80°C. On the day of use, thaw an aliquot and dilute 110 ul with 8 ml
of assay buffer. Mix gently, but thoroughly.
4.3.2.3 11-Ketotestosterone
i
This antibody was previously aliquoted and stored at -80°C. On the day of use,
thaw an aliquot and dilute 1 ul with 12 ml assay buffer. (The aliquot of antibody
can be frozen again and reused several times before starting a new aliquot). Mix
gently, but thoroughly. :
130
-------
4.3.3 Prepare Tritiated Tracer
For each steroid being measured, dilute 1 ul of the stock to 10 ml with assay buffer.
(The actual concentration is determined empirically and may change with different
batches of isotope). There should be 5000 - 6000 cpm in the 0.5 ml aliquot that is
counted from the Total tube.
4.3.4 Incubate Sample/Standards with Antibodies and Tracers
Analyze standards, controls, and if possible, the samples in duplicate. Add 100 ul of
reconstituted sample extract and standards to labeled test tubes. Add 100 ul of assay
buffer to the B0 tube and 200 ul assay buffer to Total and NSB tubes. Using the
repeater pipet, add 100 ul of diluted antibody to all tubes except the Total and NSB
tubes. Using the repeater pipet, add 100 ul of tracer solution to all tubes. Lay a sheet
of parafilm over the tubes and mix gently using the multi-tube vortexer. Incubate at
25°C for 1.5 to 2 h. If only two steroids are being measured, they can usually
incubate simultaneously, but if all three steroids are being measured, stagger the start
time of the incubation by 45 min to 1 h for one of the steroids so that conflicts due to
limited centrifuge space are avoided.
4.4 Add Charcoal Solution
At the end of the incubation, place the tubes in an ice-water bath for 15 min. Remove
the charcoal solution from the refrigerator a few minutes before needed and stir
vigorously to assure that a uniform suspension is obtained. While the charcoal solution
is still being stirred, add 400 ul to all tubes except the Total tubes using the repeater
pipet. Add 400 ul of 0.1 M PBS to Total tubes. Vortex gently and incubate again in an
ice-water bath for 15 min. Centrifuge the tubes at 3000 rpm for 30 min at 4°C.
4.5 Load Scintillation Vials
Adjust the repeater-diluter to remove 0.5 ml of the supernatant and deliver 5 ml of
scintillation cocktail. Carefully place the inlet/outlet tube below the liquid surface in the
tube and slowly remove 0.5 ml with the uptake mechanism, being careful not to disturb
the charcoal at the bottom of the tube. Then, while holding the mechanism so that it
does not release, put a scintillation vial under the inlet/outlet tube and expel the contents
that were removed from the tube. Then initiate the delivery of the scintillation fluid.
Place a cap on the vial immediately. When a whole rack of tubes has been processed,
mix by shaking.
4.7 Determine Radioactivity
Count vials in a scintillation counter for 1 min using a program appropriate for tritium.
131
-------
4.5 Troubleshooting
i
Normal results will produce a sigmoidal curve (see Section 5) with the steepest part
between approximately 0.1 and 1 ng/ml : When changing to a new source of antibody,
the dilutions may have to be adjusted to obtain the proper curve. Although the amount
of steroid in the tracer solutions can also affect the curve, a more dilute Ab will make the
assay more sensitive, shifting the curve to the left. While a more sensitive assay is
generally good, performing the assay inithis range will narrow the range of
concentrations that can be detected. Due to the normal variation between fish, having a
narrow effective range of the standards would not necessarily be advantageous. Also
making the assay more sensitive decreases the net cpm response, which will adversely
affect detection.
i
5. DATA ACQUISITION, CALCULATIONS, AND DATA REDUCTION
The cpm values from the counter are obtained on a printout and entered into a spreadsheet
(e.g., Lotus 123) for performing the necessary calculations.
5.1 Calculate B/B0
i
Calculate the B/B0 for each sample and standard. Divide the cpm value by the cpm of
the B0 sample, after the NSB value has been subtracted. Determine the B/B0 for each
duplicate well of the standards; for the samples, determine the mean B/B0 of the two
duplicates.
.
5.2 Obtain Standard Curve
Generate the sigmoidal standard curve using graphing software (e.g., SlideWrite 5). Plot
of the standards against the amount of standard present on a log scale.
5.3 Determine Corresponding Amount of Steroid in Samples from B/B0 Values
• i
Use the sample B/B0 values from the spreadsheet to determine the concentration of
steroid present in the sample.
5.4 Calculate Amount of Steroid per ml of Plasma
Transfer the extrapolated steroid values to the spreadsheet. Multiply these values by 20
(dilution factor from 5 ul of plasma present in the 100 ul of reconstituted extracted) to
give ng/ml plasma. Occasionally, an additional factor may be needed; for example, if
steroid concentrations are expected to be very high, it may be sufficient to use only 25 or
50 ul of the extracted sample rather than 100 ul.
132
-------
5.5 Determine Minimum Detection Limits
Often, particularly when measuring E2 in normal males, it may be difficult to determine
whether the result is truly zero or slightly greater than zero. If this is the case, use the
95% confidence limit (calculated with spreadsheet software) to determine if the result
should be reported as zero or as greater than zero. If the sample result is within the
confidence interval of the zero standard (NSB), report the result as zero. The minimum
detection level is the lowest standard which is consistently different than the zero
standard; that is, the two confidence intervals do not overlap. For any sample result
which is within the confidence limit of the minimum detection level or above, report the
calculated value. If the sample result falls between the zero standard and the minimum
detection level, report one-half of the minimum detection level.
5.6 Data Management and Records Management
Attach printed versions of any computer generated data in a bound record book along
with any other necessary descriptions of how the data were generated.
6. QUALITY CONTROL AND QUALITY ASSURANCE SECTION
6.1 Sigmoidal Shaped Standard Curve
The standard curve should have the typical sigmoidal appearance, flattening out at both
the low and high end of the standard range. The coefficient of determination for the
curve (determined from graphing software) should be around 0.99. If it is sufficiently
different, the plot should be examined to determine if there are potential outliers. Also,
the sensitivity should not differ from previous analyses. This would be indicated by a
change in the steeper portion of the curve relative to the standards.
6.2 LowNSBValue
The NSB value should generally be less than 400 to 600 cpm. The NSB values from a
new radioactive stock solution usually start around 100 to 200 cpm, but can increase to
600 or 800 cpm. Replace the radioactive stock solution when the value becomes too
high.
6.3 Consistent IAP Results.
The IAP result should not differ from previous analyses, or drift slowly over several
experiments.
6.4 Dependable Pipets
Check the operation of the pipets periodically for accuracy and precision. Fix or replace
any pipets not operating correctly.
133
-------
6.5 Duplicate Agreement
Duplicates agreement should be within 10%.
6.6 Ratio of B0 to the Total Counts
The maximum binding should be around 30 to 50% of the total counts in the system.
Adding less antibody will lower this value and make the assay more sensitive. Likewise,
increasing the antibody concentration will increase the B0 and decrease sensitivity.
6.7 Check of Extraction Efficiency
Check the extraction efficiency periodically by spiking a sample with a known amount
of steroid and measuring for recovery (especially if a drift in the IAP is observed).
Extraction efficiency should be at least 85 to 90%.
7. REFERENCES
Kagawa H, Takano K, and Nagahama Y. 1981. Correlation of plasma estradiol-17 p and
progesterone levels with ultrastructure and histochemistry of ovarian follicles in the white
spotted char, Salvelinus leucomaeni. Cell Tiss Res 218: 315-329.
Jensen KM, Korte JJ, Kahl MD, Pasha MS, Ankley GT. 2001. Aspects of basic
reproductive biology and endocrinology in the fathead minnow (Pimephales promelas).
Comp Biochem Physiol C 128: 127-141.
134
-------
APPENDIX D
HISTOLOGICAL TECHNIQUES FOR FATHEAD MINNOW GONADS
This appendix discusses two approaches to fixation and embedding appropriate for fathead
minnow gonadal histology: a traditional paraffin-based approach and a more modern
methacrylate-based procedure. Either technique is acceptable in the context of the test described
in this document.
1. PARAFFIN-BASED HISTOLOGICAL PROCEDURE
1.1 FIXATIVE
1.1.1 Two fixatives have generally been used for paraffin-based studies of gonadal histology and
histopathology in fathead minnows: (1) 10% neutral buffered formalin (McCormick et al. 1989;
formula example: Roberts 1978), and (2) Bouin's fluid (Roberts 1978). Presumably Bouin's is
used because it provides for rapid fixation (4-6 h) and strong subsequent tissue staining.
However, intracellular substances, such as granules and inclusions, are often poorly preserved
with this fixative (Kiernan 1990). Neutral buffered formalin gives better tissue preservation, but
the gonads should ideally be fixed for 24 h or more.
1.2 EMBEDDING ,
1.2.1 Paraffin embedding of fathead minnow gdnads can be accomplished by standard methods.
An example of this procedure, using a tissue processor, is outlined below:
70% ethanol Ih
70% ethanol Ih
80%ethanol Ih
80% ethanol Ih
95% ethanol Ih
95% ethanol Ih
100% ethanol Ih
100% ethanol Ih
Xylene 1/2 h
Xylene 1/2 h
Paraffin . Ih
Paraffin Ih
Paraffin Ih
1.3 SECTIONING AND STAINING
1.3.1 Standard sectioning and staining methods can be used with fathead minnow gonads. For
example, sections may be cut at 5 urn and stained with hematoxylin and eosin (e.g., Kiernan
135
-------
1990). Typically, the gonads are embedded longitudinally and sectioned in a step-wise fashion.
For ovaries, a number of slides are made with one or two sections from 500 um.deep into the
organ and one or two sections from 1000 UJTI deep. Testes are embedded and sectioned in a
similar manner except that the sections are taken at 250 and 500 UJOTI depths.
1.4 ADVANTAGES AND DISADVANTAGES
1.4.1 Advantages of utilizing paraffin techniques include: (1) the typical histology laboratory is
set up to routinely process large numbers of samples, (2) more personnel are familiar with
paraffin sectioning and staining procedures, and
recent EDC studies. Disadvantages include: (1)
(3) paraffin sections have been used in several
thicker sections allow less detail to be resolved,
(2) numerous artifacts (e.g., due to tissue shrinkage during processing), are present, and (3) tissue
samples have to be archived in other fixatives for subsequent high-resolution electron
microscopy.
2. JB-4 METHACRYLATE-BASED PROCEDURE
2.1 FIXATIVE
2.1.1 While the same fixatives (Bouin's, 10% neutral buffered formalin) used in paraffin
procedures may be employed, the better tissue preservation afforded by formaldehyde-
glutaraldehyde fixatives are preferable when embedding in methacrylate. Both traditional
electron microscopic fixatives, e.g., 2.5% glutaraldehyde-2% formaldehyde in 0.1M phosphate
buffer, or other fixatives, e.g., 1% glutaraldehyde-4% formaldehyde in 0.1M phosphate buffer
(Jensen et al 2001) work well with fathead minnow gonads. As with neutral buffered formalin,
gonads should be fixed for at least 24 h prior to
2.2 EMBEDDING
embedding.
2.2.1 Embedding of the relatively small fatheac
JB-4 methacrylate compared to paraffin. A typical
vials on a rotator) is as follows:
25% ethanol 30 min
50% ethanol 30 min
75% ethanol 30 min
95% ethanol 30 min
100% ethanol 30 min
JB-4 solution A (catalyzed with 6.9 g catalyst/100 ml) 2 h
Embed (40 parts catalyzed solution A: 1 part solution B; prevent contact with air
during polymerization)
minnow gonads can be accomplished rapidly in
manual schedule for gonads (with tissue in
136
-------
2.3 SECTIONING AND STAINING
2.3.1 Gonads are embedded in the same longitudinal orientation as with paraffin blocks. They
are also sectioned in the same step-wise manner as paraffin blocks, but at a thickness of 2 to 3
um. Staining of methacrylate sections can be accomplished with most stains used in paraffin
procedures, but with modifications. Examples of two modified procedures are given below.
Hematoxylin and eosin (with phloxin)
Stain sections for 30 to 45 min with filtered Harris hematoxylin
Rinse with distilled water
Dry on a hot plate
Stain cooled slides for 1 to 2 min in saturated aqueous eosin
containing 0.25% phloxin
Rinse in distilled water, dry on a hot plate, and coverslip
Basic fuschin and methylene blue-azure A
Stock basic fuschin:
1 % basic fuschin in 50 % ethanol
Stock methylene blue-azure A in distilled water:
1 % azure A
1 % methylene blue
1 % borax
Staining procedure:
Dilute basic fuschin 1:4 to 1:12 or more with distilled water
Stain 10 to 20 sec and rinse with distilled water
Dilute methylene blue-azure A 1:2 to 1:4 or more with distilled water
Stain 10 to 20 sec, rinse with distilled water, dry and coverslip
2.4 ADVANTAGES AND DISADVANTAGES
2.4.1 Advantages of JB-4 methacrylate technique include: (1) better fixation, less solvent-related
extraction, and thinner sections allow for greater resolution of tissue and cellular details
compared to paraffin sections (the superior resolution afforded by methacrylate sections
may obviate the need for electron microscopy in many cases), (2) tissues embedded in
methacrylate experience little shrinkage while those embedded in paraffin shrink 20% or
more, (3) cutting distortion, an artifact related to tissue compression during sectioning, is
greatly reduced in thin methacrylate sections, (4) staining of methacrylate sections, for
example, with hematoxylin and eosin, is far simpler than with paraffin sections, requiring
no embedment removal or sequential hydratiori and dehydration steps, (5) when electron
microscopy is required, the archived tissue is already in a suitable (recommended) fixative,
and (6) methacrylate sections have been used in several recent EDC studies. Disadvantages
include: (1) some laboratories lack the facilities for routine methacrylate procedures, and (2)
staining procedures must be modified for methacrylate sections.
137
-------
3. RECOMMENDATIONS
3.1 Where logistically possible, methacrylate embedding of tissues fixed for at least 24 h in
formaldehyde-glutaraldehyde is recommended. The greater resolution and fewer artifacts
seen in methacrylate sections make it worthwhile to establish a methacrylate facility in
histopathology laboratories where EDC studies are performed. While most stains can be
used with methacrylate, the staining procedure and stain ingredients may need to be
modified. For example, when using thin 1^3 urn methacrylate sections, the standard tissue
stain eosin may be combined with 0.1-0.5 % phloxin to improve color saturation.
4. REFERENCES
i
I
Jensen KM, Korte JJ, Kahl MD, Pasha MS, Ankley GT. 2001. Aspects of basic reproductive
biology and endocrinology in the fathead minnow (Pimephales promelas). Comp Biochem
PhysiolC 128:127-141.
Kiernan JA. 1990. Histological and Histochemical Methods: Theory and Practice. 2nd ed.
Pergamon Press, New York, NY, USA. 43J3 pp.
McCormick JM, Stokes GN, Hermanutz RO. 1989. Oocyte atresia and reproductive success in
fathead minnows (Pimephales promelas) exposed to acidified hardwater environments.
Arch Environ Contam Toxicol 18:207-2141
Roberts RJ. 1978. Fish Pathology. Bailliere Tmdall, London, England. 318pp.
138
-------
APPENDIX E
HISTOLOGICAL EVALUATION OF FATHEAD MINNOW GONADS
1. NORMAL GONADAL HISTOLOGY IN REPRODUCTIVELY- MATURE FEMALES
1.1 GENERAL STRUCTURE
1.1.1 The ovaries are paired organs that, when mature, occupy much of the abdominal cavity
ventral to the swim bladder (Grizzle 1979). They are suspended from the swim bladder by a
mesentery, the mesovarium. The mesovarium is continuous with the peritoneum that
covers the ovary proper. The peritoneum is composed of a mesothelium and an underlying
layer of connective tissue, the tunica albuginea. In fathead minnows, rodlet cells are found
among the squamous cells of the mesothelium, and eosinophilic granular cells, melanocytes,
smooth muscle, blood vessels, and nerves are common within the connective tissue layer.
The visceral border of the tunica albuginea is lined with an epithelium that is squamous in
some regions and columnar in others. The columnar cells appear to be ciliated. Ovigerous
lamellae extend from the tunica albuginea toward the center of the ovary, dividing it into
lobules that contain the oogonia and developing oocytes (Grizzle 1979).
1.2 GAMETOGENESIS
1.2.1 Fathead minnows may spawn in as little as four to five months after hatching. Little is
known about gonadal development during this time, but oogonia are identifiable in ovaries
of very young juveniles and oocytes are present in older juveniles (Grizzle 1979). However,
since fathead minnows are fractional spawners, all oocyte developmental stages are seen in
a mature ovary: (1) oogonia, (2) primary growth stage oocytes, (3) cortical alveolus stage
oocytes, and (4) early and late stage vitellogenic oocytes.
1.2.2 Oogonia (12-20 n-m diameter) , • '
1.2.2.1 Oogonia (and/or the smallest oocytes present, see Selman et al. 1993) are small cells
occurring in groups or nests along with similar-sized and larger primary oocytes (Fig. E.I;
figures follow Appendix E). Oogonia have a large nucleus with a few variably sized
nucleoli and a relatively narrow rim of cytoplasm. They are not surrounded by follicular
cells.
1.2.3 Primary growth stage oocytes (primary oocytes, 12-170 \im diameter)
1.2.3.1 Early primary growth oocytes (12-35 jjtm diameter) are often seen next to other primary
oocytes (Fig. E.I). They are still in nests and not completely surrounded by follicular cells,
which are also present in the nests. Early primary growth oocytes have a round or oval
139
-------
nucleus with a few variably-sized nucleoli. The cytoplasm contains no cortical alveoli or
yolk bodies.
1.2.3.2 Late primary growth oocytes (35-170 um diameter) have exited from cell nests and
become completely surrounded by squamous follicle cells (Fig. E.2). Both cytoplasmic and
nuclear (germinal vesicle) volumes increasp considerably during the primary growth phase,
as do the numbers of nucleoli, which tend to lie close to the nuclear envelope.
1.2.4 Cortical alvelous stage oocytes (170-425 um diameter)
1.2.4.1 Cortical alveolus oocytes are characterized by the appearance of cortical alveoli ("yolk
vesicles") and, in some species, small lipid droplets, in the cytoplasm. (It has not been
clearly established that fathead minnow oocytes have distinct lipid droplets). In early
cortical alveolus oocytes, it is possible to observe only one or two cortical alveoli; larger
oocytes have many cortical alveoli distributed throughout the cytoplasm (Figs. E.2,3). The
centrally positioned germinal vesicle is oval during this stage and has numerous peripherally
located nucleoli of various sizes. The vitelline envelope (zona radiata) is clearly visible in
even the smallest cortical alveolus oocytes; and the follicle cells are squamous in early, and
cuboidal in late, cortical aveolar oocytes.
1.2.5 Vitellogenic oocytes (425-1070 um diameter)
1.2.5.1 The initiation of vitellogenesis represents the next oocyte developmental stage which is
characterized by the accumulation of eosinophilic yolk bodies in the ooplasm. At first the
yolk bodies are much smaller than cortical alveoli and mostly dispersed among them,
especially in the perinuclear cytoplasm (Fi.g. E.3). As the oocyte grows, the yolk bodies
become larger and more numerous and displace the cortical alveoli, pushing them to the
periphery of the oocyte (Figs. E.3,6). In the late vitellogenic oocyte (800-1070 \im
diameter) the germinal vesicle also appears to move toward the periphery of the oocyte and
then disappear entirely when the oocyte approaches maturity (Fig. E.3). In vitellogenic
oocytes, the vitelline envelope thickens ancl becomes striated due to the great numbers of
pore channels that penetrate through it. The follicle cells are cuboidal with a large nucleus
and a prominent round nucleolus. External to the follicular layer lies a thin basal lamina
and a theca consisting of squamous thecal bells, capillaries, and a thin connective tissue
stroma. In the most mature oocytes observed in tissue sections the yolk bodies coalesce and
may become larger than the cortical alveoli but remain numerous rather than joining into a
single yolk mass.
1.3 STAGING OF OVARIES
i
1.3.1 Staging of ovaries of fractional spawners, such as the fathead minnow, has generally been
based upon classification of the most mature oocytes present in the histological section
(Selman and Wallace 1986; Leino et al. 1990; Selman et al 1993; Leino and McCormick
140
-------
1997; Shimizu 1997; Jensen et al. 2001). Such a classification, based on the studies above,
is presented in Table E.I. In fathead minnows, stage 5 oocytes are rarely observed unless
females are sampled while spawning. In other words, late stage 4 oocytes are apparently
only hours away from being ovulated.
Table E.I. Efistological stages of fathead minnow ovarian development.
Stage
Characteristics
1. Primary growth
2. Cortical alveolus
3. Early vitellogenic
4. Late vitellogenic
5. Mature/spawning oocyte
Oogonia and primary oocytes
la Oocytes in nests; small cytoplasmic volume
(Fig. E.I)
Ib Oocytes larger, out of nests, surrounded by
follicle cells; many pleiomorphic nucleoli
bordering the nuclear envelope (Fig. E.2)
Appearance of cortical alveoli and, possibly,
small lipid droplets (Fig. E.3)
Appearance of yolk bodies: initially few and
small; ultimately many and variably-sized;
centrally located germinal vesicle is round
to oval with several peripheral nucleoli
(Fig. E.3)
Germinal vesicle loses nucleoli, moves towards
the periphery and breaks down; yolk bodies
frequently fill the entire center of the oocyte
and a germinal vesicle may not be evident
(Fig. E.3)
Germinal vesicle breakdown complete; yolk
bodies fuse and may become larger than
cortical alveoli
1.4 OVARIAN STAGES IN NORMAL, REPRODUCING FEMALES
1.4.1 Under the test conditions described herein (25 ± 1° C, 16:8-h light:dark photoperiod) most
female fathead minnows spawn every 3 to 4 d (Jensen et al. 2001). The following
description considers variations in ovarian histology during a typical 3 d spawning cycle.
1.4.2 Day 0 post-spawn
1.4.2.1 Ovaries sampled within about 8 h after spawning (which usually takes place in the early
morning) have returned to late stage 3. Postovulatory follicles (corpora lutea) are
141
-------
numerous. In some ovaries most of these fpllicles are collapsed while in others the follicle
lumen is open (Figs. E.4,5). In goldfish, pbstovulatory follicles have lumina for at least
10 h after ovulation, but the follicles collapse by 30 h post-ovulation (Nagahama et al.
1976). This phenomenon may occur more rapidly in fathead minnows.
1.4.3 Day 1 post-spawn
i
1.4.3.1 By 1 d post-spawn, ovaries have progressed to stage 4. The post-ovulatory follicles tend
to be smaller and have thinner walls than at day 0. Some are vacuolated and appear to be
breaking down (Fig. E.6).
1.4.4 Day 2 post-spawn
1.4.4.1 By 2 d post-spawn, ovaries tend to be at late stage 4. Postovulatory follicles are often
difficult to identify. Those that are present are small and highly vacuolated.
1.4.5 Day 3 post-spawn
1.4.5.1 By 3 d post-spawn, ovaries are at stage 5 and appear spawning ready.
I
1.4.6 Atretic follicles
1.4.6.1 If an EDC affects ovarian development and/or spawning, a logical histological feature to
check for is an increase in numbers of atretic follicles (preovulatory atretic follicles,
POAFs). Follicular atresia in experimental populations of reproducing fathead minnows is
generally at a low level (Jensen et al. 2001). McCormick et al. (1989) reported a mean
atresia level (in controls) of 1.6% (range Ojl 1.6%, n=10) in their experiments, and Miles-
Richardson et al. (1999b) reported a level of 4.6% (range 0-12%, n=7). Most females have
a very low incidence of POAFs, although a small number of control specimens may have
higher levels. For example, in the McCormick et al. (1989) study, nine ovaries had aitresia
levels of 0-1.3% and one ovary had a substantially higher level of 11.6%. In a recent study,
of 27 "control" ovaries examined, only three had relatively high levels of atresia (MED,
unpublished data). Examination of these ovaries showed them to be otherwise
histologically normal, at prespawning late stage 4, with atresia mostly or entirely of the
most mature follicles (Fig. E.7). Overall, follicular atresia in the 10-12% range may be part
of a normal process in some fathead minnows some time during longer spawning periods.
2. NORMAL GONADAL HISTOLOGY IN REPRODUCTIVELY- MATURE MALES
2.1 GENERAL STRUCTURE
i
I
2.1.1 The testes are a pair of elongated white organs situated in the dorsal body cavity. Like the
ovaries they are suspended by a peritoneal mesentary. Peritoneum covers the testes and
142
-------
consists of a layer of simple squamous epithelium and a thin connective tissue capsule, the
tunica albuginea (Grizzle 1979). Connective tissue septa that separate the seminiferous
tubules are continuous with the tunica albuginea. The seminiferous tubules contain the
germinal epithelium that ultimately gives rise to spermatozoa (Fig. E.8). In mature testes
spermatozoa are present in the lumina of seminiferous tubules and of ampullae that are
similar to seminiferous tubules, but lack germinal epithilium (Grizzle 1979). Ampullae
empty into the ductus deferens.
2.2 GAMETOGENESIS
2.2.1 The germinal epithelium of fathead minnows has an apparently random distribution of
spermatogonia along the entire length of the tubule, the so-called "unrestricted" type of
testis (Grier 1981). Spermatogonia are located in small peripheral cysts in the tubule; these
cysts enlarge and extend toward the tubule lumen as spermatogenesis proceeds. Five stages
of germ cell development are readily identified in the fathead minnow: (1) primary
spermatogonia, (2) secondary spermatogonia, (3) primary spermatocytes, (4) secondary
spermatocytes, and (5) spermatids and spermatozoa (Smith 1978; Grizzle 1979; Jensen et
al 2001 and Fig. E.8).
2.2.2 In addition to the germ cells, two other principal cell-types are present in the testes: Sertoli
cells and interstitial cells of Leydig. Sertoli cell bodies, the part of the cell that contains the
nucleus, are small and difficult to locate. At high magnification these cell bodies are often
triangular-shaped structures situated near the outer rim of the seminiferous tubule (Fig. E.9).
The elongate euchromatic nucleus often exhibits a single nucleolus. Processes of a Sertoli
cell envelope a cluster of developing germ cells derived from a single primary
spermatogonium to form a cyst. These cytoplasmic processes are not usually visible with
the light microscope (Grizzle 1979). Numerous polyhedral-shaped Leydig cells are found,
usually in groups, in connective tissue spaces between seminiferous tubules (Fig. E.9).
They typically have an oval heterochromatic nucleus and a narrow rim of cytoplasm.
2.3 STAGING OF TESTES
2.3.1 Staging of the testes of fathead minnows has been based on the degree of germ cell
differentiation (Ankley et al. 2001). The presence or absence of certain stages in a
histological section, then, can be used to judge the state of testicular maturity. However, it
may be advantageous to also consider the relative size and sperm content of the
seminiferous tubules (Smith 1978; Leino et al. 1990; Gimeno et al. 1998) as in Table E.2.
143
-------
Table E.2. Histological stages of fathead minnow testicular development.
Stage
Characteristics
1. Resting germ cells
2. Spermatogonia
3. Spermatocytes
4. Spermatids and some spermatozoa in
lumen of seminiferous tubule; small
tubule lumen
5. Abundant sperm in an expanded lumen
No development
2a Primary spermatogonia:
Large cells near edges of tubule; have a
lightly staining nucleus with a prominent
nucleolus
2b Secondary spermatogonia:
Clusters of medium-sized cells with a
round, lightly basophilic nucleus; cluster
or cyst is the result of several mitotic
divisions of a primary spermatocyte.
3a Primary spermatocytes:
Smaller cells with smaller, more
basophilic nuclei than spermatogonia:
will undergo meiosis I to produce
secondary spermatocytes.
3b Secondary spermatocytes:
Small cells with smaller, more basophilic
nuclei than primary spermatocytes: will
undergo meiosis n to produce
spermatids.
Spermatids have a small, intensely basophilic
nucleus; they mature into spermatozoa
Figs. E. 10.11
2.4 TESTICULAR STAGES IN NORMAL, REPRODUCING MALES
2.4.1 During a typical 3 to 4 d spawning cycle the testes do not seem to regress to an earlier
stage as ovaries do. Examination of testes at 0,1,2, and 3 d after a spawning event indicate that,
just after spawning, certain seminiferous tubules or regions of these tubules become largely
depleted of sperm and have a thin germinal epithelium. Other tubules, however, have a thick
germinal epithelium or abundant sperm, or both. Seemingly, sperm production is unlikely to be
diminished during normal laboratory spawning.
144
-------
3. ROUTINE METHODS FOR EVALUATING EDC-INDUCED HISTOLOGICAL
CHANGES IN FATHEAD MINNOW GONADS
3.1 OVARIES
3.1.1. Gonadal staging, i.e., as described for ovaries in Table E.1, is a fundamental method for
revealing major effects of EDCs on these organs. For example, if ovaries of control fish are
stage 4, and those of EDC-exposed fish are stage 3, the EDC has produced a major effect on
ovarian development. Gonadal staging is recommended as the first step in the histological
evaluation of EDC effects. ,
3.1.2 Other methods used to describe and evaluate EDC effects on gonads are still evolving
based on the kinds of histopathological changes that are being observed, many for the first time.
Perhaps the most important of the published methods to attempt to quantify a histological
change, is counting the numbers of follicles in various stages of development (Smith 1978;
Miles-Richardson et al. 1999a,b; Jensen et al. 2001). Counting and staging individual follicles
provides information on the percentage of particular stages present. For example, treatment of
fathead minnows with 10 nM of p-estradiol for 14 d resulted in a greatly increased percentage of
primary follicles and a decreased percentage of mature follicles (Miles-Richardson et al. 1999a).
A recommended method involves counting 100 follicles from sections taken from between 500
Hm into the ovary and its midline, and calculating the percentage of each follicular stage present.
A similar method has also been employed to assess the severity of oocyte atresia (McCormick et
al. 1989). This investigation determined that a critical mean percentage of 20% atretic follicles
affected spawning success in groups of fathead minnows exposed to acidified water. Note that
this percentage is greater than the 10-12% maximal atresia occasionally seen in normally
spawning females.
3.2TESTES
3.2.1 As is the case for ovaries, testicular staging (Table E.2) represents the initial step in
evaluating the histological effects of EDCs. Also, as with ovaries, testicular staging will likely
reveal only EDC effects that profoundly influence testicular maturation.
3.2.2 Certain quantitative methods have been employed to describe more subtle changes in
testicular histology. The first involves an assessment of the percentage of each testicular stage
present, such as primary and secondary spermatogonia and spermatocytes* This information can
be used to determine whether any of the stages has an atypical distribution. Unlike with ovaries,
the relatively small and more numerous testicular germ cells are difficult to count properly
without an ocular grid or similar device. Smith (1978) employed an ocular grid to evaluate
testicular developmental stages, counting 100 cells in each of three sections per fish. It is
important to include different regions from the same testes because testicular histology
sometimes varies from one area to another, compared to ovaries which seem to have a rather
uniform distribution of oocyte stages throughout.
145
-------
3.2.3 The second quantitative method involves measurement of the tubule diameter. Certain
EDCs may enhance or decrease sperm production. Histologically this may manifest itself as a
enlargement or reduction in the mean diameter of seminiferous tubules. Smith (1978) and
Gimeno et al. (1998) described methods to measure and quantify changes in tubule diameters and
relate these changes to sperm production. Tubulp diameters should be measured in several
testicular regions, for the reasons mentioned above.
3.2.4 Special considerations for testes
I
I
3.2.4.1 Some EDO-induced histopathological changes in testes are difficult to study by light
microscopy because of the small sizes of the affected cells. This is particularly evident with
Sertoli cells which may undergo major EDC-induced changes in morphology that are difficult to
see with the light microscope (Miles-Richardson et al 1999a,b). These investigators used
electron microscopy to describe changes in Sertoli cells in conjunction with spermatocyte
necrosis. Spermatocyte necrosis appears to be aicommon result of EDC exposure (Miles-
Richardson et al 1999a,b; Ankley et al 2000; I^nge et al 2001) and electron microscopy is of
distinct benefit in supplementing light microscope-based descriptions of this pathology. Electron
microscopy may also be of benefit when studying the effects of EDCs on another small testicular
cell, the Leydig cell. Although it is reasonable to assume that EDCs may affect Leydig cells, no
studies have yet been published on this subject in fish.
3.2.4.2 Major changes in the testes, occasionally even making them difficult to identify in
histological sections, can be produced by EDCs. For example, LSnge et al (2001) reported ova-
testes and frank testicular atrophy in fathead minnows after long-term exposures to
ethinylestradiol. More work needs to be done on describing the major histopathological events
leading to testicular atrophy.
4. REFERENCES
Ankley GT, Jensen KM, Kahl MD, Korte JJ, Makynen EA. 2001. Description and evaluation of
a short-term reproduction test with the fathead minnow (Pimephales promelas). Environ Toxicol
Chem 20:1276-1290.
Gimeno S, Komen H, Jobling S, Sumpter JP, Bpwmer T. 1998. Demasculinisation of sexually
mature male common carp, Cyprinus carpio, exposed to 4-terf-pentylphenol during
spermatogenesis. Aquat Toxicol 43:93-109.
Grier HJ. 1981. Cellular organization of the testis and spermatogenesis in fish. Amer Zool
21:345-357.
Grizzle J. 1979. Anatomy and histology of the golden shiner and fathead minnow. PB-2S>4-219.
U.S. Dept. of Commerce, National Technical Information Service, Alabama Agricultural
Experiment Station, Auburn, AL, USA, 163 pp.
146
-------
Jensen KM, Korte JJ, Kahl MD, Pasha MS, Ankley GT. 2001. Aspects of basic reproductive
biology and endocrinology in the fathead minnow (Pimephales promelas). Comp Biochem
Physiol €128:127-141.
Lange R, Hutchinson TH, Croudace CP, Siegmund F, Schweinfurth H, Hampe P, Panter GH,
Sumpter JP. 2001. Effects of the synthetic estrogen 17a-ethinylestradiol on the life-cycle of the
fathead minnow (Pimephales promelas). Environ Toxicol Chem 20:1216-1227.
Leino RL, McCormick JH, Jensen KM. 1990. Multiple effects of acid and aluminum on brood
stock and progeny of fathead minnows, with emphasis on histopathology. Can J Zool 68:234-
244.
Leino RL, McCormick JH. 1997. Reproductive characteristics of the ruffe; Gymnocephalus
cernuus, in the St. Louis River estuary on western Lake Superior: a histological examination of
the ovaries over one annual cycle. Can J Fish Aquat Sci 54:256-263.
McCormick JM, Stokes GN, Hermanutz. 1989. Oocyte atresia and reproductive success in
fathead minnows (Pimephales promelas) exposed to acidified hardwater environments. Arch
Environ Contam Toxicol 18:207-214.
Miles-Richardson SR, Kramer VJ, Fitzgerald SD, Render JA, Yamini B, Barbee SJ, Giesy JP.
1999a. Effects of waterborne exposure of 17p-estradiol on secondary sex characteristics and
gonads of fathead minnows (Pimephales promelas). Aquat Toxicol 47:129-145.
Miles-Richardson SR, Pierens SL, Nichols KM, Kramer VJ, Snyder EM, Snyder SA, Render JA,
Fitzgerald SD, Giesy JP. 1999b. Effects of waterborne exposure to 4-nonylphenol and
nonylphenol ethoxylate on secondary sex characteristics and gonads of fathead minnows
(Pimephales promelas). Environ Res 80:S 122-S137.
Nagahama Y, Chan K, Hoar WS. 1976. Histochemistry and ultrastructure of pre- and post-
ovulatory follicles in the ovary of the goldfish, Carassius auratus. Can J Zool 54:1128-1139.
Selman K, Wallace RA. 1986. Gametogenesis in Fundulus heteroclitus. Amer Zool 26:173-
192.
Selman K, Wallace RA, Sarka A, Qi X. 1993. Stages of oocyte development in the zebrafish,
Brachydanio, rerio. J Morphol 218:203-224.
Shimizu A. 1997. Reproductive cycles in a reared strain of the mummichog, a daily spawner.
JFish Biol51:124-737.
Smith, RJF. 1978. Seasonal changes in the histology of the gonads and dorsal skin of the
fathead minnow, Pimephales promelas. Can J Zool 56:2103-2109.
147
-------
Figure E.I. Oogonia or small primary oocytes (arrows) in nest with larger
primary oocytes. (1170x)
Figure E.2. Late primary growth oocytes of various sizes. Late primary
oocytes are surrounded by squamous follicle cells. (380x)
148
-------
ass »wth^wv*?-Y«»»•-'--'?-
^J^§M§|;^|^!^
i %issay*iMll'.fi is, .ji
SttT,Va-«:»trt»^^/Si3SlfflH*,*4: . •*' • **•
Figure E.3. Part of ovary containing early and late vitellogenic oocytes with
densely-stained yolk bodies. Note single cortical alveolus stage oocyte
(arrow). (65x)
149
-------
Figure E.4. Post-ovulatory follicle (corpus luteum) with an open lumen
from a day 0 post-spawn ovary (see text). (190x)
Figure E.5. Collapsed corpus luteum from a day 0 post-spawn ovary. (190x)
150
-------
Figure E.6. Vacuolated corpus luteum from a day 1 post -spawn ovary. (360x)
Figure E.I. Pre-ovulatory atretic follicle. (90x)
151
-------
Figure E.8. Section of a seminiferous tubule showing various
developmental stages: (1) primary spermatogonia, (2) secondary
spermatogonia, (3) primary spermatocyte, (4) secondary spermatocyte, (5)
spermatids, late cyst above and early cyst beneath. (645x)
-------
Figure E.9. Sertoli cell (long arrow) and interstitial cells of Leydig (short
arrow) in testis. (1680x)
Figure E.10. Stage 5 testis with thick germinal epithelium and sperm-filled
lumina. (160x)
153
-------
Figure E.I 1. Stage 5 testis with thin germinal epithelium and expanded
lumina. (165x)
154
-------
-------
United States
Environmental Protection
Agency
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.
If you do not wish tojreceive these reports CHECK HEREQ;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARl
POSTAGE & FEES PAIC
EPA
PERMIT No. G-35
National Health and Environmental
Effects Research Laboratory
Duluth, MN 55804
Official Business
Penalty for Private Use
$300
EPA/600/R-01/067 •
June 2002
------- |